Final report on the participatory impact evaluation of the Root

Transcription

Root & Tuber Improvement and Marketing Program (RTIMP)
Impact Evaluation Report (June 2015)
Final Report
on the participatory impact evaluation of the
Root & Tuber Improvement & Marketing Program (RTIMP)
conducted by PDA with support from the MOFA/GOG
Pilot Application of a
Participatory Impact Assessment & Learning Approach
(PIALA)
developed with support from IFAD and the BMGF
(November 2015)
Lead authors:
Adinda Van Hemelrijck
Glowen Kyei-Mensah
Co-authors:
Nana King
Essi Haffar
Kobby Optson
Contributors:
Anthony Amuzu
Steff Deprez
Makaita Gombe
i
Impact evaluation of a the Root & Tuber Improvement and Marketing Program (RTIMP):
Funding: 1.1.1.1.1.1.1 MOFA/GOG, IFAD and the BMGF
Design & coordination:
1.1.1.1.1.1.2 Glowen Kyei-Mensah and Adinda Van Hemelrijck
Implementation:
1.1.1.1.1.1.3 PDA
Design and piloting of a Participatory Impact Assessment & Learning Approach (PIALA):
Funding: 1.1.1.1.1.1.4 IFAD and the BMGF
Management:1.1.1.1.1.1.5 Edward Heinemann, Adinda Van Hemelrijck and Richard Caldwell
Design:
1.1.1.1.1.1.6 Irene Guijt, Adinda Van Hemelrijck, Jeremy Holland and Andre Proctor
Representation and Citation:
1.1.1.1.1.1.7 The opinions expressed in this report are those of the authors and do not necessarily represent those of the
International Fund for Agricultural Development (IFAD). The designations employed and the
presentation of material in this report do not imply the expression of any opinion whatsoever on the part
of IFAD concerning the legal status of any country, territory, city or area or of its authorities, or
concerning the delimitation of its frontiers or boundaries. The designations “developed” and “developing”
countries are intended for statistical convenience and do not necessarily express a judgement about the
stage reached in the development process by a particular country or area.
1.1.1.1.1.1.8 This report or any part thereof may be reproduced for non-commercial purposes without prior permission
from IFAD, provided that the publication or extract therefrom reproduced is attributed to IFAD and the
title of this publication is stated in any publication.
1.1.1.1.1.1.9
ii
1.1.1.1.1.1.10
Table of Contents
Acronyms
List of tables and figures
List of figures
List of tables
List of textboxes
Executive Summary
Evaluation approach
Evaluation scale, scope and focus
Sampling and methodology
Training, implementation and management
Program assumptions and evaluation questions
Summary of evaluation findings and answers to the questions
Key findings regarding impact
Key findings regarding market-linking
Key findings regarding production
Key findings regarding processing
Answers to the evaluation and learning questions
Main recommendations
Main Report
Acknowledgements
Preface
1 Participatory Impact Assessment & Learning Approach (PIALA)
1.1 Evaluation approach and principles
1.2 PIALA training and management
1.3 Cost & benefits of different design options
1.4 Overview of PIALA methods and processes employed in Ghana
2 RTIMP evaluation framework
2.1 RTIMP Theory of Change (ToC)
vi
viii
viii
ix
ix
x
x
xii
xiii
xviii
xviii
xix
xix
xxi
xxii
xxiii
xxv
xxviii
1
1
3
5
5
8
9
10
14
14
2.1.1
Description of the RTIMP ToC
15
2.1.2
Visualisation of the RTIMP ToC
16
2.2 Evaluation focus
2.3 Assumptions and evaluation questions for each causal claim
3 Sampling and community mobilization
3.1 Sampling approach
18
18
23
23
3.1.1
Principle unit of analysis and sample
23
3.1.2
With/without configurations as ‘control’
24
3.1.3
Sub-samples of households and research participants
3.2
3.3
Community mobilization approach
Challenges in sampling and mobilization
24
25
26
3.3.1
Deviations in the sampling of the supply chains
26
3.3.2
Deviations in the sampling of households
26
3.3.3
Deviations in community mobilization and the sampling of research participants
27
iii
4
Background and distribution of participants and respondents
4.1 Poverty status
4.1.1
Participatory characteristics of ‘wealth & wellbeing’
29
4.1.2
Statistical categories of ‘wealth & wellbeing’
31
4.2
Household survey respondent distribution
Gender and poverty distribution in the household survey
31
4.2.2
Distribution of livelihood sources
33
34
34
35
35
35
35
36
37
37
5.6.1
Data capturing and collation
37
5.6.2
Participatory research ethics and independence
38
5.6.3
Reflective practice and quality monitoring
39
5.7 Participatory sensemaking
5.8 Methodological strengths and limitations
6 Configuration analysis
6.1 Analysis of causal links
6.2 Scoring of causal links
7 Findings on impact-related changes and causes
7.1 Changes in access to food & income and its causes (link I2I1)
39
41
43
43
43
46
46
7.1.1
Changes
46
7.1.2
Causes
48
7.2
Changes in R&T livelihoods and its causes (link O1+O2+O3I2)
50
7.2.1
Changes
50
7.2.2
Causes
50
Findings on R&T market-related changes and causes
8.1 Changes and causes of enhanced market-linking (link C1a+M1O1)
57
58
8.1.1
Changes
58
8.1.2
Causes
59
8.2 Changes and causes of supply chain linking and commercialization (link
M1c+M1b+O1+O2+O3C1b)
9
31
4.2.1
4.3 Distribution of participatory research participants
4.4 National respondents and participants
5 Field research methodology
5.1 Key Informant Interviews
5.2 Household survey
5.3 Generic change analysis
5.4 Livelihood analysis and SenseMaker lithe
5.5 Constituent feedback
5.6 Data consistency and quality monitoring
8
29
29
65
8.2.1
Changes
65
8.2.2
Causes
66
Findings on R&T production-related changes and causes
9.1 Changes and causes of enhanced R&T production (link C2a+C2bO2)
68
68
9.1.1
Changes
68
9.1.2
Causes
69
9.2 Changes and causes of access to planting materials & technologies (link
M2a+M2b+M2cC2a)
71
9.2.1
Changes
71
9.2.2
Causes
71
iv
10 Findings on R&T processing-related changes and causes
10.1 Changes and causes of enhanced R&T processing (link M3b+C3cC3bO3)
74
75
10.1.1
Changes
75
10.1.2
Causes
75
10.2 Changes and causes of access to business finance and market linking services (link
M3c+C1a+M3bC3c)
85
10.2.1
Changes
85
10.2.2
Causes
87
11 Conclusions and recommendations
11.1 Key findings related to impact
11.2 Key findings related to market-linking
11.3 Key findings related to production
11.4 Key findings related processing
11.5 Conclusions
89
89
89
90
91
92
To what extent did the assumptions hold true (or not) under which conditions?
92
What were the major barriers for farmers and processors to commercialize and access markets?
94
What is needed to make the DSF an effective mechanism for business- and market-linking?
94
What is needed to make GPCs profitable and attractive businesses?
94
What supports or hinders GPCs to better link farmers and processors to markets, and how is this
influenced by the DSF?
11.6 Recommendations
Bibliography and references
Annex 1: Correlation matrix
Annex 2: Inventory of available raw data and sub-reports
Annex 3: Overview of stakeholders interviewed
Annex 4: Distribution of participatory research participants
Annex 5: Overview of national sensemaking workshop participants
Annex 6: Sampling frame
Annex 7: Poverty distribution tables
Annex 8: Household survey questionnaire
Annex 9: Key Informant Interviews Questionnaires
Annex 10: Generic Change Analysis
Annex 11: Constituent Feedback
Annex 12: Data collation and quality monitoring
Annex 13: Participatory sensemaking at the district levels
Annex 14: Participatory sensemaking at the national level
Annex 15: Fieldwork Schedule
Annex 16: Ethical principles and standards of conduct
Annex 17: Approved Budget
94
95
99
103
105
106
109
113
116
118
120
126
133
145
150
157
162
171
175
178
v
Acronyms
ACDEP
Association of Church-based Development NGOs
AEA
Agricultural Extension Agent
ADVANCE
Ghana Agricultural Development and Value Chain Enhancement Program
AMSEC
Agricultural Mechanisation Service Centre
APFOG
Apex Farmers’ Organisation of Ghana
ARB Apex Bank
Association of Rural and Community Banks
ASAP
Adaptation for Smallholder Agriculture Programme
BAC
Business Advisory Centre
BOG
Bank of Ghana
CSIR
Council for Scientific and Industrial Research
DADU
District Agricultural Development Unit
DDA
District Director of Agriculture
DDO
District Development Officer
DFR
Department for Feeder Roads
DOC
Department of Cooperatives
DSF
District Stakeholder Fora
ERB
Enterprise Record Book
FASDEP II
Food and Agriculture Sector Development Policy
FBB
Farmer’s Business Book
FBO
Farmer Based Organization
FDA
Food and Drugs Authority
FFF
Farmer Field Forum
FGD
Focus Group Discussion
GASIP
Ghana Agriculture Sector Investment Programme
GLDB
Grains & Legumes Development Board
GLSS
Ghana Living Standards Survey
GOG
Government of Ghana
GPC
Good Practice Centers
GRATIS
Ghana Regional Appropriate Technology Industrial Service
GSA
Ghana Standards Authority
GSS
Ghana Statistical Service
ICO
IFAD Country Office
IEC
Information Education and communication
IITA
International Institute of Tropical Agriculture
KII
Key Informant Interview
MEF
Micro Enterprise Fund
MG
Matching Grant
vi
MGF
Matching Grant Facility
MLGRD
Ministry for Local Government and Rural Development
MOAP
Market Oriented Agriculture Development Programme
MOFA
Ministry of Food and Agriculture
MOTI
Ministry of Trade and Industry
NIB
National Investment Bank
NORPREP
Northern Region Poverty Reduction Programme
NRGP
Northern Rural Growth Programme
OIC
Opportunities Industrialization Centre
PCA
Principle Components Analysis
PCU
Programme Coordination Unit
PFI
Participating Financial Institution
PPMED
Policy, Planning, Monitoring and Evaluation Division
PPS
Probability Proportional to Size sampling
QCA
Qualitative Comparative Analysis
R&T
Root and Tuber
RADU
Regional Agriculture Development Unit
RAFiP
Rural & Agricultural Finance Programme
RCB
Rural and Community Bank
REP
Rural Enterprise Programme
RIMS
Results and Impact Management System
RTIMP
Root and Tuber Improvement and Marketing Programme
SCF
Supply Chain Facilitators.
SEND
Social Enterprise Development Foundation
SIFS
India Forensic Science Institute& Training Center
SME
Small and Medium Enterprise
SNV
Stichting Nederlandse Vrijwilligers / Dutch Volunteers Foundation
SRID
Statistical Research and Information Directorate of the MoFA
VCF
Value Chain Facilitator
WAAP
West African Agriculture Programme
WEAI
Women’s Empowerment in Agriculture Index
ZOC
RTIMP Zonal Office Coordinators
vii
List of tables and figures
List of figures
Figure 2.1.2
RTIMP Theory of Change diagram developed and used for this impact evaluation
17
Figure 4.2.1
% distribution of the first most important source of income for households
33
Figure 4.2.2
% distribution of the total number of sources of income for households.
33
Figure 7.1.1
% distribution of households that did not experience any food shortage in 20132014, by poverty status
47
Figure 7.1.2
% distribution of the range of total income of the household from 2009 to 2014
48
Figure 7.1.3
% distribution of total R&T production & processing value from 2009 to 2014
48
Figure 7.1.4
“The main effect of the livelihood change in my experience is...”
49
Figure 7.1.5
Correlations between food, income, R&T livelihood value and R&T Production/
Processing
49
Figure 7.2.1
“The change in my story was caused by…”
51
Figure 7.2.2
“The livelihood change in my experience was due to...”
54
Figure 9.1.1
FFF Constituent Feedback with beneficiaries – “To what extent did the training
help you expand your farming business”
73
Figure 9.1.2
FFF Constituent Feedback with beneficiaries – “To what extent have you been
able to apply what you learned at FFF”
74
Figure 10.1.1
GPC Constituent Feedback with beneficiaries – “To what extent did the training,
support and services help you to expand your processing business?”
82
Figure 10.1.2
GPC Constituent Feedback response of GPC leaders – “To what extent do you
82
think did the training, support and services provided through you GPC have
helped resource poor farmers and processors to expand their processing business?”
Figure 10.1.3
Intended beneficiaries – “To what extent have you been able to apply what you
learned at the GPC?”
83
Figure 10.1.4
Gender and Age distribution of GPC participants
84
Figure 10.1.5
GPC Constituent Feedback with beneficiaries: “How confident are you to express
your needs and ask for help at the GPC?”
84
Figure 10.1.6
Distribution of households who received finance to invest in R&T businesses in
past 5 years by poverty status
85
Figure 10.1.7
“The change in your story was influenced by...”
86
Figure 11.6.1
“Based on my experiences I think it is...”
98
viii
List of tables
Table 1.4
Summary of PIALA processes, methods & tools used in the RTIMP evaluation
11
Table 3.1.1
Sampled districts and community clusters
23
Table 4.1.1
Perceptions of wealth by gender and districts
30
Table 4.1.2
Perceptions of wellbeing by gender and districts
30
Table 4.2.1
Zonal distribution of households by gender of the household head..
33
Table 4.2.2
Distribution of households by poverty status
32
Table 4.2.3a
Poverty distribution of the Male households by Zone.
32
Table 4.2.3b
Poverty distribution of the Female households by Zone
32
Table 4.3.1
Distribution of Supply Chain Research Participants
34
Table 6.2.1
Rating of RTIMP contribution to the improvement of R&T-based livelihoods.
45
Table 10.1.1
% distribution of sources of financial support in past 5 years by range of amount
86
Table 10.1.2
Distribution of control of financial support received for investing in R&T
businesses
86
Annex 1 Table 1
Correlation Matrix
104
List of textboxes
Textbox 6.7.1
Debrief on the PIALA national sensemaking workshop (8 May 2015)
40
Textbox 7.2.1
Excerpt of farmers’ and processors’ experiences of R&T livelihood changes
51
Textbox 9.1.1
From the KIIs with the research leaders
70
ix
Executive Summary
1.
This document presents the findings from the impact evaluation of the Root & Tuber Improvement
and Marketing Program (RTIMP) in Ghana. The program was executed by the Ministry of Food and
Agriculture (MoFA), Government of Ghana (GoG) from 2007 until end of 2014, and co-financed by
the International Fund for Agricultural Development (IFAD) for a total amount of US $ 18.83
million.1
2.
Anticipating the completion of RTIMP and the start-up of the new GASIP (Ghana Agricultural
Sector Investment Program), the MoFA and the IFAD Country Office (ICO) jointly commissioned a
full-scale and -scope impact evaluation for a total of about US $ 233,000 covering the entire program
nation-wide. The evaluation was conducted by Participatory Development Associates (PDA) using a
novel Participatory Impact Assessment & Learning Approach (PIALA) developed with support from
IFAD and the Bill & Melinda Gates Foundation (BMGF). In addition to the approved evaluation
budget, the BMGF invested US $ 20,000 in methodological innovation while IFAD added about US
$ 40,000 for procurement, training, supervision and meta-inquiry of the piloting of this novel
approach PIALA as part of a broader methodological innovation project2.
3.
PIALA is not a methodology for the evaluation of performance. Hence the findings of this impact
evaluation of RTIMP do not imply a judgment on the performance of program partners and do not
question the professionalism and commitment of the Program Coordination Unit teams. Neither does
it contest the findings of the IFAD Supervision Missions and the latest Program Completion Report
about the performance and achievement of targets by the program. It offers a different perspective on
program results that is complementary to these findings: a perspective of relative influence on
changes that have impacted rural poverty, beyond the immediate effects of performance, and among
many other influences. A program, for instance, can perform well, yet have no influence, due to
various reasons that could or could not have been anticipated by the program. PIALA aims to unpack
these reasons, understand why impact occurred or not in certain circumstances, and indicate where
program mechanisms need to be revised or new ones may be needed.
Evaluation approach
4.
1
PIALA is designed to produce rigorous quantitative and qualitative evidence and generate solid
debate around such evidence in order to influence policy, planning, targeting and management for
generating greater and more sustainable impact. Its purpose is threefold: (a) to report on a project’s
or program’s contributions to impact on rural poverty; (b) to learn why impact occurred or not and
where mechanisms need to be changed or newly created; and (c) to debate how impact could be
enhanced and future program investments could have a greater influence. Different from process and
performance evaluation approaches is the focus on ‘impact’ and ‘contributions to impact’ broader
Cf. IFAD Loan No. 670, Program ID 1312. The total value of the loan was US $ 18.96 million. The total amount used of this loan
at program completion was US $ 18.83 million. The total budget used at completion was US $ 23.6 mullion, which was much lest
than the original estimated cost.
2
In October 2012, a three-year methodological innovation project was launched by IFAD in partnership with the BMGF for
developing and piloting PIALA. This was in response to a growing need for novel approaches that could help IFAD and partners
assess and understand the impacts of complex government projects and programs on rural poverty and stimulate learning.
than the intended outcomes and performance against pre-set targets. Impact is viewed from a
systemic perspective, as a system of interactions between various causes and changes, as opposed to a
more linear approach that looks at the direct relationship between intervention and effect. The
systemic approach seeks to move beyond a merely “what works” metrics and also answer the more
difficult “why” and “how” questions and investigate the likely sustainability of the changes observed.
It does so by looking at both the intended and unintended, positive and negative, primary and
secondary effects of a project or program relative to other influences that directly or indirectly
contributed to the impact on rural poverty.
5.
Generally, the questions PIALA seeks to answer are: “what has changed (or not) for whom and
why”; “how sustainable are these changes likely to be”; “what are the impacts and what has caused
these changes”; “what has been the program’s contributions to these changes among other causes”;
and finally, “what are the implications for future program strategy”. To answer all these, PIALA
draws on a systemic definition of impact, a dynamic Theory of Change (ToC) approach, participatory
mixed-methods, a participatory sensemaking model, and a configuration analysis method. This
PIALA blend of processes and methods presents an alternative for the classic counterfactual-based
evaluation in program contexts where it is quasi-impossible to find clean control groups or where
institutional and policy work has purposively “contaminated” all.
6.
In hopes of creating greater value, the PIALA processes and methods were designed3 and piloted
around three quality dimensions: rigour, inclusiveness and feasibility. Rigour is understood in terms
of methodological consistency and reliability, which in a participatory mixed-methods approach
emanates from both the rigorous employment of methods and the rigorous facilitation of
participatory processes. Acknowledging that an evaluation is never power-neutral and entirely free
from political influence or organizational pressure, and particularly not when using participatory
methods, rigor must be defined broader than in purely statistical terms and also include quality
thinking, sharp observation, engaging multiple perspectives and systematic cross-checking.4
Inclusiveness refers to the meaningful engagement of stakeholders and the credibility of findings,
requiring rigorous facilitation. Feasibility concerns the budget and capacity needed to meet the
expectations with regards to rigour and inclusiveness. A quality assurance framework (QAF) was
developed and piloted alongside the approach for assessing performance on these three dimensions in
three to four subsequent phases of the evaluation5.
7.
The approach was piloted in the impact evaluation of two IFAD-funded programs: first at a
provincial scale of the DBRP (Doing Business with the Rural Poor Project) in southern Vietnam in
2013, and subsequently at a national scale of the RTIMP in Ghana in 2015. The first pilot in Vietnam
experienced several limitations from which much was learned in the adjusted approach employed in
the second pilot in Ghana. Issues of sampling related to the heterogeneity in program distribution and
treatment, political influence and organisational pressure in the participatory processes, and
systematic data collation and quality monitoring during fieldwork to ensure data integration, were
adequately addressed in the evaluation in Ghana. This resulted in substantial improvements in the
quality of evidence.
3
The PIALA methods and tools were designed by a core team of international methods experts comprising: Adinda Van
Hemelrijck (project/team leader), Irene Guijt, Andre Procter and Jeremy Holland. Additional inputs were provided by Steff Deprez
for developing the SenseMaker tools and conducting the analysis of the SenseMaker data.
4
Cf. IFAD & BMGF, 2013c: 7.
5
The QAF is attached in the PIALA methodological reflections report. The structure of the QAF was inspired by the “Better
Evaluation Rainbow Framework” (cf. http://betterevaluation.org/plan).
xi
8.
Major strengths of this evaluation include: (a) the selection and use of methods specific to the causal
links in the ToC and the evaluation questions; (b) the comparative analysis of the relative
contribution to impact of heterogeneous configurations of program treatment (as an alternative for a
classic counterfactual analysis); and (c) the 2-stage participatory sensemaking process that engaged
all stakeholders, including beneficiaries, in a collective analysis and discussion of the evidence.
Alongside these strengths, there were also some challenges and constraints encountered by the
research team in the conduct and management of this evaluation. Three key constraints requiring
more attention in future evaluations using PIALA are: (a) the sampling of market-bounding systems
such as supply chains centred around supply chain leaders, which have per definition open
boundaries and thus are difficult to discern, particularly when interacting and thus overlapping in the
same geographic and administrative location; (b) the time and capacities required from the people to
participate in FGDs using PIALA methods, in particular when many are illiterate (e.g. the use of pen
and paper or even tablets, the length of the FGDs, SenseMaker tools using abstract concepts, etc.);
and (c) the rigid nature of the methodology that needed to be applied in a systematic manner across
all locations, which sometimes clashed with the cultural settings in some communities and was
difficult to maintain in the limited time that was spent in each district.
9.
The main take-away for future PIALA applications is that (a) methods and tools need to be adapted
to the participants’ conditions as much as possible, and (b) sufficient time is needed in the field to
accommodate cultural habits and events and address unexpected challenges with regard to sampling
and mobilisation. Obviously, if PIALA methods and tools would be used regularly as part of ongoing
M&E, then this would certainly help overcome these differences and challenges and contribute to
building participants’ capacities and empowerment. This is discussed in greater detail in a separate
report on the PIALA methodological reflections.
Evaluation scale, scope and focus
10.
Scale, scope and focus of the evaluation was agreed based on: (a) a projection of the potential costbenefits of the different design options with the commissioners before procurement was started (as
described in Section 1.3); and (b) the reconstruction of the program’s Theory of Change (ToC) with
national key stakeholders in a design workshop (as described in Paragraph § 16). The visualisation of
the ToC (cf. Figure 2.1.1 on page 13) helped identify the program’s impact and contribution claims
to be evaluated.
11.
The impact claim of the RTIMP (which is the link I2I1 in the ToC diagram) is reflected in its goal
statement, namely: “enhanced income and food security of rural poor households through
improvements in R&T-based livelihoods and strengthened market-based systems generating
profitability at all levels of the commodity chains”. At the design workshop, it was proposed to
redefine “enhanced income and food security” to avoid a too narrow interpretation of food security
as ‘food self-sufficiency’ and ensure ample attention would be paid to the profitability and
sustainability aspects of improvements in R&T livelihoods and market systems. Hence impact was
defined in terms of “access to food and income to lead and sustain an active and healthy life” and
impact-level data collection focused on essential changes in access to food & income and R&T
investments & profits.
12.
Aiming at improving rural poor people’s livelihoods in Ghana through the development of
commodity chains for Roots and Tubers (R&T) supplied by smallholders, the RTIMP consisted of
three main areas of work: a) linking of smallholders to existing and new markets; b) enhancing
smallholder R&T production; and c) enhancing smallholder R&T processing. The program design
xii
and logical framework described the causal pathway for each of these three areas through which the
program was assumed to impact on rural poverty. The evaluation needed to conduct a systemic
analysis of the interplay between these three components or contribution claims and its influences
on impact across the entire country, thus nation-wide.
13.
While the production component was started much earlier in the Roots and Tuber Improvement
Program (RTIP)6 that preceded the RTIMP, interventions related to enterprise upgrading and marketlinking were added under RTIMP, some of which became effective on a national scale only after the
2010 Mid-Term Review (MTR). Hence the main reference period for evaluating the interplay
between the three components concerned the last 5 years of program implementation, starting at the
start of 2010 (or at MTR). The 2008 RIMS baseline was used for comparison of findings only with
regard to production.
14.
The evaluation focused on the four main types of commodity chains developed during this period,
namely: Gari, High Quality Cassava Flour (HQCF), Plywood Cassava Flour (PCF) and Fresh Yam
for Export (FYE). Furthermore, the focus was on the four main program mechanisms that would be
considered for scaling up in the new GASIP –namely: the District Stakeholder Forum (DSF), the
Farmer Field Forum (FFF), the Good Practice Centre (GPC) and the Micro-Enterprise Fund (MEF).
The evaluation serves to flag emerging issues from the RTIMP that merit closer attention in the
GASIP, more innovative thinking, and more evaluative input, and therefore was framed as a learning
exercise that complements other M&E and supervision processes.
Sampling and methodology
15.
The catchment or ‘supply chain’ areas of the commodity chains formed the principle unit of analysis
for inquiring the interplay between the three RTIMP components and its influences on impact.
Supply chains consist of ‘supply chain leaders’ (such as gari and HQCF producing GPCs and
factories, plywood factories and fresh yam exporters) and ‘suppliers’ (smallholder producers and
processors), and are geographically defined by their location. Since the supply chains were
administratively served at the district level, 25 districts were randomly sampled from the 67 districts
treated by RTIMP7 at the time of the evaluation design across the main 3 agro-ecological and
administrative zones. The 25 districts comprised 30 community clusters, each comprising 3
communities and locating a supply chain. The 30 community clusters contained samples of supply
chains of the 4 commodities with probability proportional to seize (PPS) of their total populations of
supply chains. Some deviations (discussed in Section 3.3.1 of the main report) occurred in the supply
chain samples though, which reduced the amount of researched supply chains from 30 to 25 (largely
corresponding to the 25 districts). This made the fieldwork and data collation more onerous, but
didn’t affect the quality of the evidence. The deviations are discussed in Section 3.3.1.
16.
Sufficient coverage of heterogeneity in program treatment was ensured by including all the different
with/without configurations of the evaluated program mechanisms in the sample. The sample also
contained several districts where the mechanisms were mostly dysfunctional or not in place, which
formed a useful comparison group that provided ‘counterfactual’ evidence at the level of the
observed R&T livelihood changes (not at the household level).
6
The RTIP focused primarily on cassava research and development. The RTIMP extended this focus to other roots and tubers and
added a strong marketing component designed to improve poor farmers' access to food and income.
7
The programme’s completion report says that the programme had worked in 106 districts across all ten regions by the end of its
operational period. At the time of the evaluation design though a list of 68 treated distrocts was provided by the program
coordination unit for sampling.
xiii
17.
With RTIMP effects spilling over and many other rural livelihoods programs influencing rural
people’s lives and livelihoods all over Ghana, it was very difficult (if not impossible) to find nonconfounded or non-influenced communities and households that could serve as control groups for
inquiring the net attributable impacts of RTIMP on household poverty. There was also no interest
among the core learning group (established with the sponsors and key stakeholders at the design
workshop in Kumasi on 12 October 20148) to collect evidence from household-level control groups
at the cost of a systemic inquiry of supply chains. Hence it was decided not to conduct a classic
counterfactual inquiry of rural poverty impact at the household level, but instead to conduct a
configuration analysis of the effects of different ‘with/without’ configurations of program
mechanisms on changes in R&T livelihoods that impacted household food and income. The
evaluation was framed as a learning exercise and thus sought to understand the explanations for their
contributions in terms of reach, effectiveness and sustainability.
18.
To assess changes in household food & income, and the influences of changes in R&T livelihoods on
these, a brief household survey was conducted. For this, 30 households were randomly sampled in
each of the 30 sampled community clusters, by systematically selecting every 10th or 5th household
starting from the central community centre. Although the supply chains were reduced from 30 to 25,
the original sample of 30 community clusters was upheld for subsampling the households, as to
ensure the total sample size would be sufficient to arrive at 95 % statistical precision. In 2 clusters in
the Kumasi Metropolitan Assembly, no suppliers could be found and thus no households sample as
‘intended beneficiaries’ (cf. Section 3.3.2 in the main report). Moreover, three household surveys
could not be accounted for, which brought the total amount of surveys down from 900 to 837 (184 in
the Northern, 424 in the Central and 229 in the Southern zone).
19.
Also the participatory research participants were subsampled in the original sample of 30 community
clusters, minus the 2 clusters in the Kumasi Metropolitan Assembly where no ‘intended beneficiaries’
could be found. Average 42 intended beneficiaries were selected in each of the 28 community
clusters (in total 1180), using an 80/20 ratio of primary to secondary beneficiaries and a 50/50 gender
ratio with 10-20% young adults (<35 years). In principle, the participants were selected separately
from the household survey respondents and quasi-randomly from a list of beneficiaries obtained from
the district officials or by using a snowballing technique where no lists were available. Due to some
deviations though there was some overlapping though between the survey respondents and the
participatory research participants in 4 districts (described in Section 3.3.2 of the main report).
20.
The participatory research involved Focus Group Discussions (FGDs) using three different types of
participatory data collection methods to further investigate the causes of the observed changes in
R&T livelihoods that affected household food and income, and the contributions that the evaluated
program mechanisms made in the area of production, processing and market linking. In total, 109
gender-specific FGDs (53 women and 56 men), in which 839 community members participated (411
women and 428 men; and 90 % intended beneficiaries) and 43 gender-mixed FGDs with a total of
341 participants (179 women and 162 men) were held. The methods were selected specific to the
causal links in the program ToC, namely9:
8
The participants in this workshop were invited to further take part in the evaluation’s Core Learning Partnership (CLP). These
included: the RTIMP Coordination Unit and Steering Committee, the IFAD Country Program Office ,the MoFA, PFIs, RIs and
SCFs and TREND. The major outcome of the evaluation design workshop was the design paper (cf. Van Hemelrijck, A. & G.
Kyei-Mensah, 2014). Also report on the workshop proceedings was produced separately.
9
Both the Constituent Feedback (cf. http://www.keystoneaccountability.org/analysis/constituency) and the SenseMaker (cf.
http://cognitive-edge.com/sensemaker) were methodological experiments funded by the BMGF. There were limited in size and
ambition and merely served the purpose to pilot-test their complementarity to the other PIALA methods and their added value for
impact evaluation. The findings on this are presented in a separate report on the PIALA methodological reflections.
xiv

The generic change analysis, which is a PRA-inspired method that combines two tools (a
change ranking and a causal flow mapping of changes in wealth & wellbeing) to further
investigate the impact claim, in addition to the household survey;

The livelihood analysis, which is a method that combines two PRA-inspired tools (change
matrix and causal flow mapping) and a small SenseMaker exercise to investigate R&T
livelihood changes and causes;

The Constituent Feedback, which collects quantified perceptual data on the reach and effects
of the program mechanisms (DSF, FFF and GPC/MEF) on R&T livelihood changes and causes.
21.
Additionally, over 100 Key Informant Interviews (KIIs) were conducted, of which 75 with districtlevel and over 25 with regional and national program stakeholders. At the regional and national level
these included RTIMP and IFAD officials, managers from the PFIs, the FFF research leaders, and a
few important off-takers or industry leaders. At the district-level these were district officials, leaders
of GPCs and other SMEs, and the managers of the local branches of the PFIs.
22.
Special tools were also designed and used for early (almost instant) data linking and quality
monitoring during fieldwork. This made it possible to organise debates with local stakeholders
around the emerging evidence in district-level sensemaking workshops immediately after collecting
the data in each district, and also ensured the evidence that was built would be robust enough to
permit causal inference and stand up to scrutiny.
23.
A 2-stage participatory sensemaking process was organised that engaged all stakeholders in a
collective analysis and discussion of the evidence in relation to the links in the ToC. For this a
workshop model has been developed and pilot-tested in Vietnam and Ghana. Half-day sensemaking
workshops were organized in 23 of the 25 districts, engaging 640 research participants (average of 28
per workshop), of which 81 % intended beneficiaries (48 % female and 52 % male farmers and
processors, mostly illiterate). A two-day national sensemaking workshop was organised on 6-7 May
in Kumasi, involving 106 participants, of which 40 % intended beneficiaries (38 % female and 62 %
male farmers and processors, many illiterate), 45 % local and national officials, and 15 % private
sector actors (including bankers and service-providers). All the participants in the district and the
national sensemaking workshops were purposively sampled from the research participants in the
field research and the KIIs. Through these multi-stage sensemaking processes, local and national
stakeholders were actively engaged in a collective analysis and debate of the evidence of RTIMP
contributions to livelihood improvements and impact on rural poverty.
24.
Last, a novel configuration analysis method was developed for the impact evaluation of RTIMP
that enabled clustering and comparing a large amount of evidence across the 25 districts to surface
the patterns of interaction and influence in/between the different intervention areas (or contribution
claims) of the program, and draw general conclusions with regard to program contributions to impact.
This was done by first ‘unzipping’ the theory of change from the impact level to the level of the
program mechanisms, in order to identify clusters of districts presenting different combinations of
program treatment, outcomes and conditions for each of the causal claims in the ToC, and
subsequently ‘zipping up’ the findings again along the ToC, in order to draw conclusions about
program contributions to impact. Combining QCA10 techniques with a systemic ToC approach, the
10
Qualitative Comparative Analysis (QCA) is a methodology used for analysing large and small n data sets by identifying all
possible combinations of variables observed in the data set, and then applying the rules of logical inference to determine which
conclusions are supported by the data. In the case of the evaluation of RTIMP, logical inference was applied to different
combinations of program treatment (or the functioning of the program mechanisms), outcomes (reflected in the scores of relative
strength and consistency for each of the causal links in the ToC) and conditions (described in the qualitative evidence) in order to
xv
method offers an alternative way to arrive at rigorous causal inference in the absence of clean control
groups. This is particularly useful for programs/projects where it is quasi-impossible to find such
clean control groups, or where institutional and policy work has purposively ‘contaminated’ all. The
configuration analysis method is presented in Section 5.
25.
The table below presents an overview of the PIALA methods and processes employed in the
evaluation of RTIMP in Ghana, and the participants that took part in each of these. The total net
amount of participants without overlap was over 2000 (incl. 837 households, 1180 FGD
participants with some overlap with the households in 4 districts, and over 100 KII participants).
PROCESSES, METHODS & TOOLS
PURPOSES
PARTICIPANTS
I. DESIGN: Focusing and framing the evaluation
Projection of potential cost-benefits of
different design options (Section 1.3)
 Enable commissioners to make a
decision about scale, scope and purpose
of the evaluation based on an adequate
Methods/tools:
understanding of the different design
Outline of 3 design options (full scale–
options in terms of quality, outcomes
full scope; limited scale–full scope; full
and budget implications
scale–limited scope) in relation to the 3
PIALA purposes (reporting, advocacy,
learning)
 the IFAD Country Program
Manager
 the RTIMP Coordinator
representing the
MoFA/GoG
 Identify the program’s impact and
contribution claims to be evaluated, and
formulate evaluation questions focused
on these claims and their assumptions
Methods/tools:

Create a shared understanding of the
Emerging ToC diagram that shows the
program’s ToC (including broader
envisioned causal pathways (with
influences on impact)
codification of the causal links) elicited

Select the methods specifically in
from the program documents and the
relation to the causal links in the impact
discussions with national stakeholders\
and contribution claims
 National key stakeholders
who had been involved in
program implementation,
management and
supervision (total of 32
participants –incl. RTIMP,
IFAD, MoFA, PFIs, RIs &
SCFs)
Reconstruction and visualisation of the
program’s Theory of Change (ToC)
(Section 2.1)
II. FIELDWORK: Collecting and linking the data
 Enable a systemic inquiry of the impact
of the combined changes in production,
processing and market linking on
livelihoods and poverty status in 30
Methods/tools:
random supply chains across the country
 Sampling hierarchy
 Enable comparative analysis of the
 Data collection & methods table
systemic inquiries of the 30 supply
 ‘How-to’ guidance for employing
chains
the data collection, collation and
 Ensure rigorous employment of methods
quality monitoring methods
and facilitation of participatory
 Standard note-taking formats
processes
 Ensure systematic data capturing, data
collation, data quality monitoring and
reflective practice during fieldwork
Data collection on changes in access to  Collect and triangulate data on the link
food & income and its causes
I2I1 in the ToC
(Sections 5.2 & 5.3)
 Engage beneficiaries of RTIMP in a
discussion of changes in livelihoods
Methods/tools:
affecting household wealth and
 Household survey
wellbeing, based on a visual
 Generic change analysis (incl.
reconstruction of the actual causal
change ranking and causal flow
pathways
Sampling and developing methods and
tools for data collection, collation and
quality monitoring (Sections 2.3 & 3.1)
 PDA research team (incl.
research assistants), GSS
statistician, 2 methods
consultants
 837 households (random)
were surveyed
 439 intended program
beneficiaries (quasi-random;
51 % women and 49 %
men) participated in the
generic change analysis
arrive at solid conclusions about the program’s influences on livelihood improvements and poverty status. More information about
QCA can be found on: http://www.u.arizona.edu/~cragin/fsQCA/index.shtml.
xvi
mapping of changes in wealth &
wellbeing)
Data collection on changes in R&T
 Collect and triangulate data on the link
livelihoods and its causes (Section 5.4)
O1+O2+O3I2 in the ToC

Engage beneficiaries of RTIMP in a
Methods/tools:
discussion of changes in production,
 Generic change analysis (see
processing and market linking affecting
above)
their livelihoods, based on the visual
 Livelihood analysis method (incl.
reconstruction of the actual causal
change matrix exercise, causal
pathways
flow mapping, and SenseMaker
Data collection on reach and effects of
selected program mechanisms (DSF,
FFF, GPC/MEF) (Sections 5.5 & 5.1)
Methods/tools:
 Livelihood analysis (see above)
 Constituent Feedback (using a
specific set of facilitation and
scoring questions for each
mechanism)
 Semi-structured interviews
(mirroring the scoring questions
in the Constituent Feedback)
Data consistency and quality
monitoring (Section 5.6)



Methods/tools:
 Standard data collation table

 Daily team reflections using five
standard sets of questions (use of

methods, facilitation of processes,
data capturing, sufficiency of data
on causal links, and sufficiency of
data on program mechanisms)
 400 intended program
beneficiaries (quasi-random;
47 % women and 53 %
men) participated in the
livelihood change analysis,
of which 393 did the
SenseMaker exercise
(participants in the generic
change analysis see above)
Collect and triangulate data on the causal  341 beneficiaries
links between the program mechanisms
participated in the
(DSF, FFF & GPC/MEF) and the
Constituent Feedback (53 %
observed changes in production,
women, 47 % men)
processing and market linking (O1, O2
 100 officials and service
& O3)
providers (75 district-level
Engage beneficiaries of RTIMP in a
and 25 regional/national)
group discussion and anonymous scoring
participated in the KIIs
of the reach and benefits of the services
(participants in the livelihood
provided through the program, and the
change analysis see above)
effects of these on the changes in
production, processing and market
linking that affected their livelihoods
Identify data gaps and weaknesses early  PDA field research teams
on in fieldwork to enable researchers to
 supervision by the PDA
probe for more information in the
research coordinator and the
sensemaking workshops
IFAD consultant
Ensure evidence is robust (inclusive,
statistically rigorous)
Instant data processing and crosschecking during fieldwork making it
possible to organise debates with local
stakeholders around emerging evidence
in district sensemaking workshops
III. ANALYSIS: Synthesizing the evidence and analysing and debating program contributions
Participatory sensemaking
(Section 5.7)
Processes:
1. half-day local sensemaking
workshops in 23 of the 25
sampled districts
2. 2-day national sensemaking
workshop
Methods:
 reverse engineering
 active listening
 patches & nodes
 iterative & recursive design
 vantage points
 soft systems modelling
 contribution analysis
Configuration analysis (Section 6)
Methods/tools:
 aggregated data collation table
 configuration analysis method
 Obtain additional information and fill in
remaining data gaps
 Help program stakeholders develop a
more systemic understanding of the
development processes impacting rural
poverty
 Engage program stakeholders in
discussing and valuing program
contributions to rural poverty impact,
and identifying priority areas for
investment
 Give voice to those who were intended
to benefit, while offering decisionmakers and service-providers the
opportunity to engage in dialogue with
these voices, based on evidence
 640 local research
participants in district
(81 % intended beneficiaries
of which 48 % women and
52 % men)
 106 local, regional and
national participants in
national sensemaking
workshop (40 % intended
beneficiaries of which 38 %
women and 62 % men;
45 % officials; 15 % private
sector actors)
Participants were selected from
the field research participants
(with exception of households)
 Arrive at rigorous causal inference in the  IFAD consultant (PIALA
absence of clean control groups
project leader)
 PDA research coordinator
and research team leaders
xvii
Training, implementation and management
26.
The design of the evaluation and the training of the research team was headed by Adinda Van
Hemelrijck (IFAD consultant / PIALA project and team leader) while the management, coordination
and field supervision was led by Glowen Kyei-Mensah (PDA managing director). The two worked
closely together in partnership, thus both contributing and taking joint responsibility for the design,
quality and results of the evaluation.
27.
Design and training took place from mid-October until mid-December 2014, and involved two days
of PIALA design training, one week11 of desk review and reconstruction of ToC, one week of
stakeholder consultations and design workshop, one week of methods training and tool development,
and three days of field-testing. The products that came out of this process included the evaluation
design paper12 and a researchers’ handbook based on the PIALA guidance provided by the PIALA
design team. During this period, improvements were made to the methodology in response to the
feedback received from IFAD and external reviewers on the first PIALA pilot in Vietnam. The
methods and tools used in Vietnam were revised and new methods and tools were developed
specifically for the impact evaluation in Ghana. This was all done as part of the design and fieldtesting.
28.
The evaluation was conducted by three teams each consisting of four researchers speaking multiple
local languages: one team per zone (North, Central and South). Each team was divided in two subteams of two working in parallel. Field research was undertaken during six weeks (from early
January until mid-February 2015) in twenty-five districts in eight regions across the country. An
average of 4-5 days was spent in each district for mobilising research participants, collecting data and
organising a sensemaking workshop. Fieldwork was supervised by Glowen Kyei-Mensah and
Adinda Van Hemelrijck who alternately accompanied the teams in the North, Central and South.
Program assumptions and evaluation questions
29.
RTIMP’s Theory of Change consists of one impact claim and three contribution claims.13 The impact
claim is reflected in the program’s goal statement and consists of two causal links: (1) the link I2I1
in which R&T livelihood improvements creates greater access to food and income for the rural poor,
and (2) the link O1+O2+O3I2 in which enhanced market linking, production and processing
realize the R&T livelihood improvements. The contribution claims reflect the RTIMP’s three
intervention areas through which it sought to realize these outcomes: market-linking, production and
processing.
30.
The assumptions of these causal claims that needed to be inquired were:

With regard to impact:
o livelihoods and poverty status could be improved by commercializing smallholder R&T
production and processing businesses and developing competitive market-driven and
inclusive supply chains;

With regard to market linking:
11
A week counts for 5 days of work.
Cf. Van Hemelrijck, A. & G. Kyei-Mensah (2014). Design Paper for the impact evaluation of the Root & Tuber Improvement &
Marketing Program (RTIMP). Participatory Impact Assessment & Learning Approach (PIALA) developed with support of IFAD
and BMGF.
13
See also Paragraph § 11 of the Executive summary, and Sections 2.1.1 and 2.1.2.
12
xviii
o
o
DSFs would help develop sustainable and inclusive R&T commodity chains;
more recourse-poor R&T farmers and processors (including women and young adults)
would commercialize and become part of the supply chains, if they would obtain the
knowledge and capacity to increase their production, access markets and develop viable
businesses;

With regard to production:
o FFFs would enable resource-poor R&T farmers and seed producers to become
commercial growers by organising into FBOs and adopting improved planting materials
and technologies;

With regard to processing:
o well-trained processors and farmers would be able to obtain a loan through the MEF to
invest in their businesses;
o GPCs would reach and teach resource-poor farmers and processors about good quality
processing practices and the use of improved technologies and standardized equipment,
hence helping them access the MEF and develop profitable businesses.14
31.
The evaluation questions were:
 to what extent these assumptions held true (or not) and under which conditions; and
 what the major barriers were for farmers and processors to commercialize and access old and
new markets.
32.
In addition, insights were sought in relation to the following learning questions:
 what is needed to make the DSF an effective mechanism for business- and market-linking;
 what is needed to make GPCs profitable and attractive businesses; and
 what supports or hinders GPCs to better link farmers and processors to old and new markets and
how is this influenced by the DSF.
Summary of evaluation findings and answers to the questions
33.
A summary of the key findings from the aggregated analysis of the evidence collected on each causal
claim and each causal link is presented here in reverse order, starting from the changes and causes at
the impact level and ending with the effects of the evaluated program mechanisms (DSF, FFF, GPC
and MEF). Based on these findings, answers are formulated to the evaluation questions for the
program’s impact claim and three contribution claims (market-linking, production and processing).
Key findings regarding impact
34.
In terms of impact on rural people’s access to food and income (cf. the link I2I1 in the ToC
diagram, presented by Figure 2.1.2 on page 13) in the RTIMP treatment areas, the evidence shows
three major trends that have occurred in past 5 years (2009-2014).
14
Although all assumptions were extensively discussed, reviewed and approved at the evaluation design workshop (before
fieldwork began), RTIMP officials who had participated in this workshop explained at the national sensemaking workshop (after
the field work was finished) that in general it was not the responsibility of the GPC to reach and teach farmers and processors.
Amendments to the assumptions however should have been made in the design workshop. Moreover, IFAD funding targets rural
poverty by enhancing small farmers’ and processors’ ability to develop businesses and access markets, and thus one would expect
that the funding for the upgrading of enterprises into GPCs should contribute one way or another to the development of these small
businesses by exposing farmers and processors to good practices and providing them with access to improved technologies and
equipment. The extent to which this was realised has been inquired by this evaluation as ‘reach’.
xix
35.

First, there has been an increase in access to food and income among rural households. However,
R&T livelihood changes did not predominantly affect access to food but rather access to income.
This has been confirmed by evidence from two independent sources. The pattern analysis of 373
micro-narratives collected from the livelihood change FGDs showed that 93.5 % of the farmers
and processors consider these changes as having an impact mainly on income. The statistical
analysis of the 837 household surveys showed a more direct thus linear correlation of
households’ total value of R&T production/processing with total household income (p.54;
sig.000) than with households’ access to food (p.22; sig.000).

Second, 15 % of the households have raised their income above USD 2/day between 2009 and
2015, which largely can be attributed to improvements in R&T livelihoods, and thus can be
considered as a positive impact. However, when looking at the percentages of households who
invested in R&T production (50 %) or processing (11 %) in the past 5 years, as compared to the
relatively small percentage of households (10 %) that gained value up to USD 2-4/day from
R&T production and processing, and the zero amount of households (-1 %) that moved into
higher R&T livelihood values above USD 4/day, it is clear that the impact has remained limited
and unsustainable. These figures came out from the statistical analysis of the household
surveys, while explanations were produced by the ‘livelihood change’ and ‘generic change’
FGDs, for which the evidence was found robust and consistent across all 25 researched districts.

Third, as more households moved into R&T farming and consequently production volumes
increased, investments in R&T farming decreased while also access to technologies decreased,
partially due to a shift from production to processing. Investments and profits from enhanced
R&T production and processing remained limited though, and livelihood improvements lingered
fragile. Again, quantitative figures from the statistical analysis were confirmed and explained by
the evidence from the ‘livelihood change’ and the ‘generic change’ FGDs, which was found
robust and consistent across the country.
In terms of changes in livelihoods and the influence of enhanced market linking, production and
processing on these changes (cf. the link O1+O1+O3I2 in the ToC diagram, presented by Figure
2.1.2 on page 13), four major findings came out from the aggregated analysis of all the evidence
collected from district KIIs, ‘livelihood change’ mappings and constituent feedback scorings in the
25 researched districts. The evidence was found quite strong and consistent across all the districts.

First, in 52 % of the researched supply chains15, improvements of R&T-based livelihoods
between 2009 and 2015 were found relatively strong though not all attributable to RTIMP. The
other 48 % generally performed weak in this regard. In 32-33 % of the researched supply chains,
positive as well as negative livelihood changes were clearly attributable to RTIMP.16 Overall,
its strongest contribution was made in the area of production; it’s weakest in the area of market
linking17. In a few cases (e.g. North Dayi/Kpando, Agona East, Pru, Tano North/Dua Yaw
Nkwanta, and Wassa Amenfi West), RTIMP mechanisms were dysfunctional or not in place,
hence providing counterfactual evidence for the difference that the program has made in the
supply chains where changes were found clearly attributable to RTIMP. In these few cases,
15
Originally, 30 community clusters in 25 districts were randomly sampled (each comprising 3 communities and locating a supply
chain) with probability proportional to seize (PPS) of the total populations of supply chains of the four researched commodities.
Some deviations occurred in the supply chain samples though, which limited the actual amount of researched supply chains from 30
to 25 (corresponding the 25 districts). See also Paragraph § 15 of the Executive summary.
16
This came out from the configuration analysis as well as the SenseMaker analysis (cf. Paragraphs § 142-144 in the main report).
17
RTIMP had three intervention areas through which it sought to affect R&T livelihoods and household poverty status: marketlinking, production and processing. See also Paragraph § 28 of the Executive summary.
xx
livelihoods improvements were very weak or virtually absent and barely influenced by RTIMP
or other programs.

Second, FFFs undoubtedly made a positive difference in 84 % of the supply chains. R&T
production boomed across the country largely due to the introduction of new seed varieties and
farming technologies by the RTIMP and its predecessor RTIP. The new varieties increased the
value and volume of raw and processed R&T produce and contributed to the increase in
household income. This initially caused an influx of people into R&T farming, which led to a
substantial increase of production volumes and triggered a spill over into processing.

Third, markets largely failed to absorb the increasing production volumes, which turned the tide
and caused prices to drop, hence negatively affecting farmers’ and processors’ livelihoods from
2013 onward. Accelerated by the economic downturn18, inadequate market linking due to weak
DSF performance in 84 % of the cases hampered the growth of farmers’ and processors’ profits
and investments, rendering improvements in their livelihoods fragile.
Key findings regarding market-linking
36.
Regarding the changes in market-linking and the influence of the DSFs on these changes (cf. the link
C1a+M1O1 in the ToC diagram on page 13), the following four key findings came out from the
aggregated analysis of the evidence collected from the KIIs, the FGDs on ‘livelihood change’ and the
constituent feedback scorings on DSF performance in the 25 districts. The evidence was found
generally quite strong and in most cases fairly consistent.

Market linking of supply chains through the DSF was found weak and ineffective in more than
84 % of the researched districts across the country. In 57 % of these, DSFs to some extent
contributed to strengthening the supply chains, but largely failed to link the supply chains to
sufficient markets. In 43 % of the cases, DSF contribution to developing the supply chains was
virtually nil and no efforts were made to link farmers and processors to markets.

Where supply chain development and livelihood improvements were found relatively strong,
despite weak market linking, this was due to a stronger performance of other RTIMP
mechanisms (in particular GPCs and FFFs) and the influence of other organisations. Where
livelihood improvements were weaker, generally RTIMP and other organisations had a weaker
presence and consequently negative trends such as high inflation and dropping prices
exaggerated by poor infrastructure had a bigger impact on income levels. Where also the supply
chains were weak, resource-poor farmers and processors were much more vulnerable to unfair
competition/trade and power abuse by clan leaders and middlemen controlling the farm-gate
prices and the gate to the local markets.

Only in 16 % of the researched supply chains19, supply chain development and the attraction of
new buyers was comparatively more effective, which enabled farmers and processors to expand
their businesses. This was largely due to a stronger performance of DSFs and GPCs. Both
mechanisms were instrumental in developing supply chains and linking these to new markets.
Yet they have not proven strong enough to withstand external threads and prevent market
saturation, due to the GPC’s insufficient capacity to innovate and expand, which was further
18
Ghana experienced an economic crisis since 2013, with the Ghana cedi dropping up to 40% against the US dollar in 2014. Cf.
The Guardian (8 August 2014), Reuters (13 May 2015), The Economist (20 June 2015).
19
Originally, 30 community clusters in 25 districts were randomly sampled (each comprising 3 communities and locating a supply
chain). Due some deviations in the field, the actual amount of researched supply chains were reduced from 30 to 25 (corresponding
the 25 districts). See also Paragraph § 15 of the Executive summary.
xxi
constraint by the licensing requirements of the Food & Drugs Authority (FDA) and the Ghana
Standards Authority (GSA).
37.
Regarding the commercialisation and supply chain linking resulting from enhanced production,
processing and market-linking (cf. the link M1c+M1b+O3+O2 +O1C1b), there are four major
findings.

The first is that R&T production and processing has changed across the entire country from a
merely food producing subsistence to a commercial income-generating livelihood. Both
household surveys and KIIs have confirmed this trend. Evidence collected from districts not
treated by RTIMP (such as in Agona East, Pru and Wassa Amenfi West) showed the necessity of
supply chain development and market linking for enabling smallholders to commercialise. In
the absence of any intervention in this area, resource-poor farmers remain extremely vulnerable
to unfair competition/trade.

Second, commercialisation has remained limited and unsustainable in more than 88 % of the
researched districts across the country largely due to market saturation as a result of weak and
ineffective market linking combined with overproduction. Poor roads and poor market
infrastructure further limited resource-poor farmers’ and processors’ market opportunities and
in the absence of appropriate competition regulations rendered them more vulnerable to unfair
competition/trade including monopolistic behaviour of GPCs.

Third, in 12 % of the supply chains, commercialisation was found to be relatively stronger but
inconsistent and not entirely attributable to RTIMP. In these cases GPCs (as supply chain
leaders) have proved to be an important mechanism to make it possible for resource-poor farmers
and processors (as suppliers) to develop small profitable businesses and gradually grow and
commercialize. The success of this mechanism was largely due to its capacity to innovate and
create new market value/demand, its reach of farmers and processors in the catchment area,
and the trust it has built between the various supply chain actors and their buyers and service
providers to establish strong supply chains.

Last, while FFFs have been very successful in enhancing smallholder production by introducing
improved planting materials and technologies in more than 84 % of the supply chains, they have
proved insufficient to enable farmers to organise and commercialise. Although useful for
various donor programs to better target and reach farmers, FBOs have not proven sufficient to
enable farmers to better defend themselves against unfair competition/trade and power abuse,
gain better access to finance and markets, and increase profits and investments.
Key findings regarding production
38.
Regarding enhanced production and productivity due the adoption of improved planting materials
and technologies and farmers’ organisation (cf. the link C2a+C3b O2 in the ToC diagram on page
13), 2 key findings came out from the evidence from the KIIs, the FGDs and the constituent feedback
scorings on in the 25 districts. The evidence was found generally quite strong and consistent (score 46).

First, R&T production and productivity increased substantially in about 76 % of the researched
supply chains due to the adoption of improved planting materials and technologies. The
substantial increase though has caused a saturation of local markets, which hampered farmers’
profits and investments and their ability to further commercialise. Where results were rather
weak, this was due to a limited adoption as well as other influences such as beetle infestation,
xxii
changing weather patterns, limited markets, land tenure issues, and a limited affordability of
inputs.

39.
Second, there is no evidence that supports the idea of FBOs as having been an effective
mechanism for helping farmers bargain better prices, fight unfair competition, obtain business
finance, access markets and commercialize. All evidence points to the need for more market
opportunities (and thus better roads and market places, and policies and regulations more
supportive of smallholder business development) to enable farmers to commercialise in the first
place, and to the need for developing mixed agri-business organisations that are less centred
around merely farming and more attuned to value creation (thus including agri-processing and
market-linking activities).
As for the changes in resource-poor farmers’ and seed producers’ access to improved planting
materials and technologies due to the FFFs (cf. the link M2a+M2b+M2cC2a), there are three
important findings:

First, FFFs have proven an effective mechanism to promote the adoption of new planting
technologies and seed varieties, because of their highly participatory character. Widespread
adoption was mainly due to the unsurpassed efficacy of the planting in rows using appropriate
distances and agrochemical application technologies, and the visible benefits in terms of a
substantial increase in quantity/quality and value (in particular for cassava).

Second, although women are generally more involved in cassava production than men and
traditionally do most of the work, FFFs mostly targeted and reached men, in particular smallscale male farmers between 40 and 60 years old who own a bit of land (max 2 ha). Since R&T
changed from a food crop to a cash crop, men took a greater interest and FFFs have encouraged
and supported this. As a result FFFs insufficiently reached and supported women.

Third, most FFF beneficiaries reported that they were able to apply what they learned at FFFs,
which helped them expand their businesses, but young farmers (< 25 years) and women were less
positive than adult men, and also felt less confident to express their needs and ask for help at
FFFs. Since in most tribes, women don’t talk or participate in FFFs, women-specific FFFs should
have been organised.
Key findings regarding processing
40.
Regarding changes in processing volumes and quality resulting from changes in farmers’ and
processors’ access to improved technologies and equipment and their ability to develop profitable
businesses supported by the GPCs and the MEF (cf. the link M3b+C3cC3bO3 in the ToC
diagram on page 13), three major findings came out of the configuration analysis of the evidence
obtained from the 18 districts with gari and HQCF supply chains (which are the only 2 commodities
that involve processing). The evidence came from the KIIs, livelihood change FGDs and constituent
feedback scoring on GPC performance, and was overall found quite strong.

First, processed volumes of cassava increased considerably in about 50 % of the gari and HQCF
chains (or 9 out of 18) as a result of more people processing cassava and expanding their
businesses by gaining access to training and facilities at GPCs. In only 3 of the 9 districts (all
gari supply chains), this was found fairly robust and attributable to RTIMP due to stronger
performing GPCs in terms of market creation, reach of farmers and processors, and the
development of stronger and more inclusive supply chains. Adoption and use of improved
technologies and equipment through the GPC was quite high. In the other 6 cases where
xxiii
enhanced processing was found strong but inconsistent, this was caused by GPC operations with
a more limited reach on the one hand, and by the spill-over of excess production into
processing20 that used both new and traditional equipment on the other. Farmers started to
process their excess cassava into gari but mostly in the old fashioned way, yet were able to
produce more and better quality gari due to the new cassava variety. In the 50 % supply chains
(or 9 cases) where enhanced processing was found weak, this was due the very limited reach of
the GPCs (more than half of which were not functional) and the very limited use/adoption of
improved technologies and equipment by resource-poor farmers and processors (which was
found nearly nil in more than half of these cases).
41.

Second, where improved processing technologies and standardized equipment were
effectively used, generally processing volumes and quality increased significantly. Access to
these technologies and equipment was created by introducing a cassava processing equipment
package, training local artisans to manufacture improved agro-processing equipment and provide
repair and maintenance services, and by upgrading small processing enterprises to GPCs that
could serve as demonstration, learning and practice centres and as market-hubs for processors
and farmers. As the cost-benefit analysis of processing equipment conducted in 2014 clearly
showed, the new technologies and equipment have proven cost-efficient and attractive in terms
of their potential return on investments (MoFA, 2014a).

Third, adoption/use of improved processing technologies and standardized equipment have
proven ineffective in 15 of the 18 sampled cassava processing districts due to the limited reach
and effectiveness of GPC’s as learning and good practice centres and the limited investment
capital of small processing centres and individual farmers and processors. Both the household
survey and the FGDs undeniably showed limited profits and investments in R&T-based
livelihoods and limited access to financial support to invest in existing or new livelihood
activities. Farmers and processors attributed negative livelihood changes mainly to the lack of
access to finance. Of those reached by GPCs (mostly women, average 35 % < 35 years), nearly
one third found that these had helped them expand their businesses, and over half stated they
were able to apply what they had learned at the GPC, thus showing the relevance of GPCs.
Women were generally more positive and less neutral than men, although they appeared less
confident to speak out, express their needs and ask for help at GPCs. Also people younger than
25 felt less confident. Interestingly, only 9 % of the GPC leaders were positive about the
influence of the GPC on smallholder business development and people’s ability to apply what
they had learned.
Regarding the changes in access to business finance for investing in improved processing
technologies with support from the MEF (cf. the link M3c+C1a+M3bC3c in the ToC diagram on
page 13, there is one major conclusion, for which strong evidence was drawn from the household
survey, the KIIs and the livelihood analyses (including SenseMaker).

The MEF was not available and accessible to the majority of farmers and processors as well as
to most GPCs, hence did not make a noticeable difference to their adoption of improved
technologies and equipment and the development of their businesses. The mechanism was
formally unavailable in more than half of the districts. Only in a few cases was there evidence of
groups of processors and farmers that obtained finance through the MEF or other channels to
20
Also the correlation analysis of the household survey data clearly indicated such a shift from production to processing as a result
of excess production (cf. Paragraph § 126 in Section 7.1.2).
xxiv
invest in their businesses. There were 3 important reasons for this: (i) the procedure 21 for
obtaining and paying off MEF funding appeared too onerous, making smallholders pre-invest
and sustain operations without sufficient capital or immediate returns on their investment;
(ii) PFIs showed reluctant to approve applications because of the perceived risk of investing in
farming and agri-processing businesses; and (iii) the present conjuncture made R&T smallholder
business investments too precarious for financing.
Answers to the evaluation and learning questions
42.
The findings of this evaluation leads to the conclusion that, R&T-based livelihoods initially did
improve between 2009 and 2013, which was relatively strong in about half of the districts and
affected income levels with 15 % of households raising above the threshold of USD 2/day. This
influenced households’ access to food. However, these positive impacts remained limited and
unsustainable largely due to market insufficiency starting from 2013. This was particularly so in
those districts where supply chains and DSF performance was found weak and inadequate, and also
where GPCs did not take up any role in the supply chain linking of small farmers and processors and
did not contribute to the development of their businesses. In these districts, market insufficiency
combined with an inadequate rural infrastructure and land tenure system negatively affected small
and resource-poor farmers’ and processors’ livelihoods and poverty status from 2013 onward, when
the economic downturn struck the country. Evidence points at a reasonable attribution of positive as
well as negative livelihood changes to RTIMP in 32-33 % of the supply chains and of farmers’ and
processors’ individual experiences (cf. Paragraph § 34 of this Executive Summary).
43.
The assumption that livelihoods and poverty status could be improved by commercializing
smallholder R&T production and processing businesses, and by developing competitive and
inclusive supply chains, only held true where very strong and concerted efforts were made by the
program partners to: (a) develop solid links between the supply chain actors; (b) address their
capacity and relational issues; (c) create new market opportunities; and (d) expand the catchment
area by widening and deepening the reach of resource-poor farmers and processors in the
surrounding communities. In particular, where the performance of DSFs and GPCs in this regard
were the strongest (12 %), supply chain development and commercialization was more successful,
resulting in greater livelihood improvements. Where the performance of these mechanisms were
weak, investments in smallholder businesses remained limited and profits stayed in the hands of a
few, thus undermining the hypothesis of smallholder commercialization as the driving force for
sustainable livelihood improvement and poverty reduction. However, also in those few districts with
better DSF and GPC performance, livelihood improvements remained fragile due to insufficient
capacity on the part of the GPCs to innovate and expand, further constrained by FDA and GSA
licensing requirements, export regulations, border taxes, and the failing power supply and
infrastructure.
21
The MEF procedure was the following: The DADU undertook a needs assessment on the prospective beneficiaries as a basis for
possible financing. Upon submission of an application (mostly ranging between 728 GHS to 60,000 GHS), the PFI then inquired if
the potential beneficiary met the requirements. In the case of the Ecobank for instance, processors were required to submit firm
orders with pro-forma invoices and contracts from key customers before loan approval. If the potential beneficiary met all criteria,
then the loan was approved and the application was sent to the RTIMP national office that then granted authorization to transfer the
matching grant component to the requested PFI. Finally, a supplier was paid to manufacture and deliver the requested equipment to
the MEF beneficiary after s/he fulfilled his/her 10 % contribution to the investment.
xxv
44.
The assumption that DSFs would help develop sustainable and inclusive R&T commodity chains
largely did not hold true. In 84 % of the sampled districts, DSFs failed to help link farmers and
processors to markets, and in 43 % of these also failed to help establish sustainable and inclusive
supply chains. Lacking were the resources and capacities at the districts (and the higher support
levels) to make this mechanism work –e.g. to conduct proper market analysis and integrated VC
development planning, attract private investment, promote product diversification/innovation,
support market creation for smallholder businesses, deepen and expand the reach and role of the DSF,
and propose legislative and policy changes at higher levels needed to make actions at local levels
more successful.
45.
The assumption that more resource-poor R&T farmers and processors (including women and young
adults) would commercialize and become part of the supply chains, if they would be able to increase
their production, access markets and develop viable businesses, only held true in the few cases
where these conditions were fulfilled by strong GPC and DSF performance. Generally, limited
commercialization and ineffective supply chain linking was largely due to:




the limited reach and capacity of DSFs and GPCs to expand, innovate and develop markets;
unfair competition and monopolistic behaviour by traders, entrepreneurs (including GPCs) and
popular leaders (including of MoFA officials);
lack of market opportunities due to a failing rural infrastructure and inadequate policy and
regulations supportive and protective of smallholder business development (including unfair
competition, licencing and certification, export and border tax, etc.); and
lack of trust and investment capital of resource-poor farmers and processors for the above
reasons. All these causes together hindered resource-poor farmers’ and processors’ ability to commercialize
and enter new markets, and thus outweighed the initial benefits from enhanced R&T production and
processing.
46.
The assumption that FFFs would enable R&T farmers and seed producers commercialise by
organising into FBOs and adopting improved planting materials and technologies has proven
partially true. FFFs undoubtedly made a positive difference in 84 % of the supply chains due to
farmers’ massive adoption of the new varieties and technologies, which increased the value and
volume of raw and processed R&T produce and contributed to the increase in household income.
However, counterfactual evidence showed the necessity of strong supply chains and market links to
enable farmers to commercialise. Also no evidence was found that FBOs could be an effective
mechanism for helping farmers bargain better prices, fight unfair competition, obtain business
finance, access markets and commercialize. All evidence points to the need for agri-business
organisations that are less centred on merely farming and are more attuned to market value creation.
47.
The assumption that GPCs would reach and teach resource-poor farmers and processors to use
improved technologies & equipment, access business finance and develop profitable businesses, held
true only in a few cases where GPCs more deliberately took up this role (thus functioning more as
social-private profit) and had a greater capacity. Adoption and use of improved technologies and
standardized equipment by resource-poor processors has proven limited in 15 of the 18 sampled gari
and cassava flour supply chains, due to the limited reach and effectiveness of GPC’s as learning
and good practice centres and the limited investment capital of small processing centres and
individual farmers and processors.
48.
Finally, the assumption that well-trained processors and farmers would be able to obtain a loan
through the MEF to invest in their businesses by large has proven untrue. Resource poor farmers
xxvi
and processors were unable to access MEF as the mechanism was formally unavailable in over half
of the sampled districts and mostly inaccessible in the other half due to the risks involved.
What were the major barriers for farmers and processors to commercialize and access markets?
49.
The two most fundamental barriers that are conditional for addressing all other limitations are:
 Lack of market opportunities due to a failing infrastructure (in particular power supply, water,
roads and market places);
 Lack of investment capital (only 15 % of the households obtained some sort of financial support
for investing in their R&T businesses in the past 5 years; 45 % of the negative livelihood
experiences were attributed to the lack of access to finance).
50.
Another important limitation is the lack of capacity of farmers and processors to organise into
independent and collective agri-businesses that are able to create market value. A more conducive
policy environment and rural infrastructure, however, are conditional to this.
51.
From the findings, it came out clearly that more resources and capacities at district and regional
levels are needed to:
 conduct market analysis and develop plans for integrated VC development;
 attract investments for transportation and infrastructure development;
 promote product diversification/innovation and support market/demand creation among GPCs
and other small enterprises with sufficient outreach in the VC catchment areas;
 organise more regular DSF meetings that are open to all supply chain actors and accessible to
more remote communities for discussing market opportunities and issues of unfair competition;
 undertake appropriate action to address the issues raised at DSF meetings and propose changes
in policy and regulations at higher levels needed to make actions at local levels more successful.
52.
In the 3 cases where livelihood improvements were found strongest, GPCs were essential to make it
possible for processors to develop profitable business and gradually commercialize. The success of
this mechanism was largely due to its capacity to innovate and create new market value/demand, its
reach of farmers and processors in the catchment area, and the trust it built between the various
supply chain actors and their buyers and service providers to establish strong supply chains.
53.
Of those reached by GPCs (mostly women, average 35 % < 35 years), nearly one third found that
these had helped them expand their businesses, and over half stated they were able to apply what
they had learned at the GPC, thus showing the relevance of GPCs to resource-poor farmers and
processors (including women and young adults).
What supports or hinders GPCs to better link farmers and processors to markets, and how is
this influenced by the DSF?
54.
Most essential impediments identified by this evaluation include:





Limited operational capital
Limited capacity to innovate and expand
Failing power and water supply
Expensive licensing and certification procedures
Rising export and border taxes
xxvii




Rising transportation costs
Limited reach of farmers and processors
Private profit orientation centred on elite interests
Monopolistic behaviour (e.g. unfair price setting, breach of agreements, etc.)
Main recommendations
55.
It must be clear by now that RTIMP has made an substantial contribution to the development of
opportunities for resource-poor farmers and processors to improve their lives and livelihoods by
turning R&T (the most important crops grown by the majority of people in Ghana) from a merely
subsistence into a cash crop. There is no doubt that this very important shift is largely attributable to
RTIMP. Plenty of evidence has been provided by this impact evaluation that supports this
conclusion.
56.
Having acknowledged this important step forward, there is also the need now for a more sobering
reflection on the factors and conditions that have hampered the sustainability of the positive impact
that RTIMP has generated on the lives and livelihoods of the rural poor in Ghana. Although FFFs
have proven a very effective mechanism, its success has resulted in excess production that saturated
local markets in almost all districts, hampering farmers’ profits and investments and their ability to
further commercialise. Evidence from districts not treated by RTIMP has sufficiently proven the
necessity of strong supply chains and market links to enable resource-poor farmers and processors to
commercialise. Without sufficient markets, impacts from enhanced production and processing are
unsustainable.
57.
Our first critical reflection and recommendation concerns the highly successful FFF mechanism.
Intuitively, everyone would recommend a scaling-up of this mechanism –with adjustments to ensure
greater gender and generation sensitiveness, e.g. by organizing gender- and youth-specific groups.
This definitely would contribute not only to enhancing the value of R&T production, but also to
women’s and young farmers’ empowerment. Being an important source of creativity and innovation
that have remained largely untapped, women and youth (<25 years) would definitely benefit form
their organization into business-oriented farming and agri-processing groups. The FFF concept might
be a suitable mechanism to explore and unlash this idea. However we must inquire and carefully
monitor the conditions that are essential to make this ‘idea’ successful in a conjuncture of rising
inflation and failing markets. Hence we recommend the piloting and scaling up of the formation of
gender- and youth-specific groups with very careful monitoring of the conditions required to avoid
harm to their livelihoods and trigger the successful growth of these groups into small collective agribusinesses.
58.
RTIMP performance was generally weak in the area of market linking. Clearly, there is the urgent
need to rethink the DSF mechanism. Commonly DSFs were organized around the supply chain
leaders, mostly small and medium-sized agri-processing enterprises that were turned into GPCs. In
doing so, its reach was limited to the farmers and processors in these specific supply chains, making
them dependent on the supply chain leaders’ benevolence, thus providing the leaders free rein to
monopolize the supply chains and the local markets. The DSF should become a forum that supports
inclusive supply chain linking and encourages innovation and diversification in value creation. By
doing so it can provide room for all farmers and processors and engage them in multiple short and
long value chains. Also traders, transporters, bulkers and off takers need to take part in DSF
meetings. Sufficient resources and capacities at the districts (and the higher support levels) are
needed to make this mechanism work –e.g. to conduct proper market analysis and integrated VC
xxviii
support market creation for smallholder businesses, reach out for farmers and processors and
particularly for women and youth to engage them in the development of strong value chain linkages,
and propose changes in policy and regulations needed to create market opportunities and protect
farmers and processors from unfair competition.
59.
GPCs were crucial to make it possible for processors to develop profitable business and gradually
commercialize. The one-third of processors (mostly women, average 35 % < 35 years) that expressed
their satisfaction with the functioning of the GPCs and the benefits they gained has clearly shown its
relevance to resource-poor farmers and processors (particularly women and young adults). However
the success of the GPC mechanism was limited as it was unclear what is required to be an effective
‘leader’ in developing strong and inclusive supply chains. The potential power of strong business
relationships was shown in a few cases where GPCs functioned as open social-private profit centres
where resource-poor farmers and processors learned to use improved technologies and equipment
and create added value of their produce. Where GPCs were profitable and attractive businesses in
particular for women and young processors, this was largely due to its capacity to innovate and
demonstrate innovation and thus create new market value/demand, its reach of farmers and
processors in the surrounding communities, and the trust it built between the various supply chain
actors and their buyers and service providers to establish strong and inclusive supply chains. Hence
our recommendation here is to expand the concept of GPCs, properly define its leadership role, and
use appropriate performance and feedback monitoring criteria and tools that help keeping track of the
quality and effectiveness of its business relationship with farmers and processors in the surrounding
communities (in particular those resource-poor). Moreover, similar as for the FFF/FBO-mechanism,
careful monitoring is required of the conditions under which GPCs can become effective
leaders of strong and inclusive supply chains.
60.
To help GPCs and FFF/FBOs as small collective agri-businesses build their capital and
investments, there is an urgent need for feasible finance mechanisms. Commercialisation and
adoption of improved technologies has remained limited in most of the supply chains, not only due to
a limited reach and capacity of GPCs and DSFs, but mostly due to the lack of finance and market
opportunities. The MEF mechanism attempted to address the issue of finance, yet has largely failed.
Its procedure for obtaining and paying off the loan made it difficult for beneficiaries to pre-invest and
sustain operations without sufficient capital or immediate returns on their investment. Hence the
mechanism needs to be completely restructured in order for it to be accessible to small farmer and
processor businesses. Repayment periods and requirements need to be feasible and agreed upfront.
More thought need to be put in into ‘risk transfer’ and ‘risk distribution’ mechanisms and criteria for
credit worthiness, and into developing different credit packages targeting different categories of
businesses involved in the VCs. Finally, there should be a more comprehensive consultation and
communication process to make all actors involved in the VCs fully understand the risks, the
mechanisms and the requirements regarding repayment and investment. However, to make any credit
mechanism work, feasible business and market opportunities must exist, which in many places in
rural Ghana currently don’t exist.
61.
In order to give all these mechanisms a reasonable chance to succeed and sustain, much more work is
needed on creating an environment that is more conducive of the growth of small collective agribusinesses. Essential is a minimal rural infrastructure (roads, market places, power and water
supply), which in many places in Ghana is failing. Crucial are also policies and regulations with
regard to fair competition and the use of cassava in end products. A policy that compels industries
producing flour, starch, beer or bio-fuels in Ghana, for instance, to include a percentage of cassava
xxix
flour in their products, would definitely spur the demand for cassava. Second, a policy and authority
that regulates competition to make it fair and inclusive, protect smallholder businesses, and prevent
monopolistic practices would certainly aid the DSF and GPC mechanisms to build stronger business
and market relationships between the value chain actors and stimulate greater inclusion of small
farmers and processors. (ODI, 2010)
xxx
Main Report
Acknowledgements
Not everyone is comfortable with an impact evaluation piloting a quite innovative and unconventional
approach in its own backyard, and particularly not with a full scope of the entire program under evaluation
and at a full national scale. The impact evaluation of RTIMP was the second PIALA pilot and the first that
was conducted at a national scale, which was only possible due to the endless trust, commitment and
patience of the IFAD Country Office (ICO) in Ghana, in particular its Country Program Manager Ulac
Demirag, and of the RTIMP Coordination Unit, in particular its program coordinator Adjekum Akwasi.
Most essential was of course also the trust and support of the MoFA Chief Director ING. Maurice Tanco
Adisa-Seidu expressed in his approval of funding of the impact evaluation for a total of US $ 232,827.
Active support to the procurement and supervision of the evaluation was also received from the IFAD
Country Program Officer Theophilus Otchere Larbi and Financial Manager Daniel Passos.
This report is the result of a multi-layered team effort that started when in September 2012 Ed Heinemann
(Lead Technical Advisor on Policy at the Policy & Technical Advisory Division at IFAD) and Adinda Van
Hemelrijck (evaluation consultant and project leader for IFAD) suggested the idea of developing a
Participatory Impact Assessment and Learning Approach (PIALA) for IFAD. They hired a team of
international consultants to help design and pilot PIALA and enabled funding through IFAD’s Innovation
Mainstreaming Initiative and a partnership with the BMGF. Richard Caldwell from the BMGF (head of
Measurement, Learning and Evaluation in the Agricultural Development Program) invested much trust by
agreeing to co-fund and co-manage the project. Ed, Richard and Adinda together managed the overall
PIALA effort and played a key role in generating wide interest and commitment at every step of the
piloting process. The total funding put in by IFAD and BMGF for the design, supervision and meta-inquiry
of the second PIALA pilot in Ghana, and for the testing of some new methods such as Constituent
Feedback and SenseMaker, counted for about US $ 70,000.
The core design team of international consultants –including Irene Guijt, Andre Proctor, Jeremy
Holland, and Adinda Van Hemelrijck – developed the first PIALA concept22 into a workable pilot at a
provincial scale in Vietnam23. Building on the essential learnings from the methodological reflections on
this first pilot24, the concept was further expanded to a national scale for the second pilot in the impact
evaluation of RTIMP in Ghana. A group of key people at the different departments at IFAD actively
engaged in the reflections and provided their critical feedback on the first pilot in Vietnam and the scalingup concept for the second pilot in Ghana25. Special thanks in this regard goes to Clare Bishop (Lead
Technical Specialist on Gender and Inclusion at the Policy & Technical Advisory Division), and Fabrizio
Felloni (Lead Evaluation Officer at the Independent Evaluation Office). Furthermore, the second PIALA
pilot also benefited from the thoughtful feedback on the first Vietnam pilot from an external reference
22
Cf. IFAD & BMGF. (2013c). PIALA Research Strategy. Improved Learning Initiative.
The impact evaluation in Vietnam was conducted of the IFAD-funded DBRP in Ben Tre province (Mekong delta). The total cost
of the evaluation was about US $ 90,000 with an additional US $ 100,000 for the PIALA design, supervision and reflections).
24
Cf. IFAD & BMGF. (2013b). Improved Learning Initiative for the design of a Participatory Impact Assessment & Learning
Approach (PIALA): Insights and lessons learned from the reflections on the PIALA piloting in Vietnam.
25
Cf. IFAD & BMGF (2013a). Potential scalability of PIALA for impact M&E &L. Piloting options and budget considerations for
Ghana.
23
1
group of independent experts –in particular from Kent Glenzer (MIIS), Robert Chambers (IDS), and
Marie Gaarder (IEG at the World Bank).
The design of the impact evaluation of RTIMP was headed jointly by Adinda Van Hemelrijck (evaluation
consultant and PIALA project leader for IFAD) and Glowen Kyei-Mensah (PDA’s Managing Director and
the research coordinator for the RTIMP evaluation). The sampling and improvement of the quantitative
methodology, and the analysis of household survey data, was headed by Anthony Amuzu (senior
statistician at the Ghana Statistical Service). Important inputs were provided for improving the mixedmethods package, in particular the participatory methods and configuration analysis method, by Irene
Guijt and Andre Proctor from the PIALA core design team. Last special gratitude goes to them for
remaining closely engaged in this second PIALA pilot. Additional inputs were obtained from Steff Deprez
for designing and piloting the SenseMaker tools and conducting the trend analysis of the micro-narratives,
and from Makaita Combe for the analysis of the constituent feedback data. Also the feedback on the
design of the RTIMP evaluation from a few people at IFAD in Rome, in particular Edward Heinemann and
Fabrizio Felloni, and the third member in the PIALA core design team, Jeremy Holland, was received with
much gratitude.
The data on which this report is based was collected with the dedication and critical reflection of the PDA
research team, coordinated by Glowen Kyei-Mensah (Managing Director) with support from Nana King
(Program Manager) and Helen Nti (Administration and Finance Manager). Special gratitude goes to these
three managers for their perseverance, flexibility and trust despite the many challenges and anxieties this
pilot has caused their organisation and researchers. Data collection and data collation as well as all the
sensemaking workshops were skilfully led and facilitated by a core team of four highly enthusiast and
studious lead researchers –including Essi Haffar, Kobby Optson, Nana King and Bernard Alando.
Junior researchers who assisted the lead researchers with the field research logistics and the data collection,
were: Usif Osman Wuntuma, Abubakari Abdu Samed, Beatrice Sarpong, Natasha Botchway,
Mohammed Abubakari, Gina Ama Gyan, Bismark Dzahene-Quarshie, Elikem Aggor, Isaac Quansah
and Kwadwo Anokye. Finally, support to the report writing was also provided by Jonathan Anaglo
(lecturer at the University of Ghana).
2
Preface
The past ten years have seen a surge in interest, practice and investment in impact evaluation in
international development. There is an increasing demand for impact evaluation to assist policy-makers and
donors in understanding and enhancing the effectiveness of public investment. Bulletproof numbers are
required to justify program investments at scale, while credible explanations are needed of the observed
impacts to influence national policy and local responsibility for greater and more sustainable impact. Large
investment programs however are increasingly complex and political, with many actors involved having
different stakes, making it difficult to find a one-size-fits-all methodology that can respond to all the
different knowledge needs. This complexity is likely to increase in the context of the new SDGs as
demands for greater inclusiveness and sustainability are added to those of effectiveness, and impact
evaluations are expected to contribute to building inclusive, responsible and sustainable societies by
enabling citizens to critically engage with evidence of impact. Hence there is a growing need for more
complexity-sensitive systemic and inclusive approaches for impact evaluation that employ participatory and
mixed quantitative and qualitative methods to assess the impact trajectories of these increasingly complex
and politicized investments.
Acknowledging this need and keen to better understand impact in ways that go beyond a simple ‘what
works’ metrics, IFAD and the BMGF co-funded a 3-year innovation project for the design and piloting of a
novel approach called PIALA to assess and explain the impacts of IFAD-funded government programs on
rural poverty in a collaborative and participatory manner. PIALA has come a long way after three years of
intensive collaboration with the various stakeholders in and outside IFAD to develop its piloting strategy,
conduct its first pilot in Vietnam, critically reflect upon the results from this first pilot, and conduct the
second pilot in Ghana using an improved and scaled version of the approach. The admirable audacity of the
Government of Ghana (GoG), and more specifically of its Ministry of Food and Agriculture (MoFA), to opt
for a nation-wide and full-scope impact evaluation of its program on agricultural smallholder and market
development for one of Ghana’s most prevalent and important cash and foot crops, namely roots and tubers,
offered a welcome opportunity to further expand and pilot-test PIALA.
This report conveys the findings from the impact evaluation of the Root & Tuber Improvement and
Marketing Program (RTIMP) in Ghana that piloted PIALA. The evaluation was conducted by Participatory
Development Associates (PDA) with support and supervision from IFAD. The program was executed by
the MoFA/GoG from 2007 until end of 2014. IFAD had co-financed for 59 % of its total budget26. The
MoFA/GoG and the IFAD Country Office (ICO) in Ghana jointly commissioned the impact evaluation
upon completion of RTIMP and in anticipation of the next Ghana Agricultural Sector Investment Program
(GASIP). The evaluation serves to flag emerging issues from the RTIMP that merit closer attention in the
GASIP, more innovative thinking, and more evaluative input, and therefore was framed as a learning
exercise that complements other M&E and supervision processes in three fundamentally different ways:
26

it offers an independent perspective and thus a critical sounding board, as it collected descriptive and
explanatory evidence of changes and causes across a representative random sample of the supply
chains for the four commodities that were inquired (gari, HQCF, fresh yam for export, and PCF);

it also offers a systemic perspective of relative program contribution to livelihood changes impacting
rural poverty, beyond the immediate effects of performance and among other influences; and
Cf. IFAD Loan No. 670, Program ID 1312. The total value of the loan was US $ 18.96 million.
3

it also gives voice to those intended to benefit from the program and offers officials, serviceproviders and private sector actors involved in the program a unique opportunity to listen to and
engage in dialogue with these voices, based on evidence.
PIALA is not a methodology for the evaluation of performance. Hence the findings of this impact
evaluation of RTIMP do not imply a judgment on the performance of program partners and do not question
the professionalism and commitment of the Program Coordination Unit teams. Neither does it contest the
findings of the IFAD Supervision Missions and the latest Program Completion Report about the
performance and achievement of targets by the program. If offers a different perspective on program results
that is complementary to these findings: a systemic perspective of relative influence on changes that have
impacted rural poverty, beyond the immediate effects of performance. A program for instance can perform
well, yet have no influence, due to various reasons that could or could not have been anticipated by the
program. PIALA aims to unpack these reasons, understand why impact occurred or not in certain
circumstances, and indicate where program mechanisms need to be revised or new ones may be needed that
have not been tried yet. Hence ‘performance’ of program mechanisms is looked at in a different way: rather
than measuring achievement of pre-defined targets, it permits to rigorous assessment of relative program
contributions to observed livelihood changes and make a sound judgement of its impact on poverty.
Knowledge-sharing, finally, is an integral and essential part of this piloting initiative, not at least because of
its emphasis on critical reflection and learning by program and evaluation stakeholders about the trade-offs
and cost-benefits of different design decisions in different program and evaluation contexts, but also
because of the novelty of PIALA and its potential use for IFAD, for the GoG, and for the wider sector.
Many good ideas are coming out of the piloting and sharing with the wider sector to further explore and
expand the boundaries of what’s cutting edge in participatory theory-based mixed-methods impact
evaluation. This is largely thanks to the commissioning and funding of this ambitious impact evaluation by
the MoFA/GoG and ICO Ghana!
August 2015
4
1
Participatory Impact Assessment & Learning Approach
(PIALA)
1.1 Evaluation approach and principles
1.
IFAD has to report to its Members States on the total number of rural people lifted out of poverty27.
Given its multidimensional systemic approach and the complex and politicized nature of the projects
and programs it funds, this is quite challenging. The question for IFAD is not just about how to
rigorously measure its attributable share, but also about how to reach such ambitious targets through
working with the governments it funds and the partners who co-implement and co-fund the
programs.28 PIALA is designed to produce rigorous quantitative and qualitative evidence and
generate solid debate around such evidence in order to influence policy and planning for greater and
more sustainable impact. Its purpose is threefold: (a) to report on a project’s or program’s
contributions to impact on rural poverty; (b) to learn why impact occurred or not and where
mechanisms need to be changed or newly created; and (c) to debate how impact could be enhanced
and future program investments could have a greater influence.
2.
Although initially piloted in ex-post impact evaluation, PIALA is thought of as an approach for
collective impact monitoring and evaluation throughout a program’s lifetime. Different from
process and performance evaluation approaches, however, is its focus on ‘impact’ and ‘contributions
to impact’ broader than the intended outcomes and performance against pre-set targets. Impact is
viewed from a systemic perspective, as a system of interactions between various causes and changes,
as opposed to a more linear approach that looks at the direct relationship between intervention and
effect. The systemic approach seeks to move beyond a merely “what works” metrics and also answer
the more difficult “why” and “how” questions and investigate the likely sustainability of the changes
observed. It does so by looking at both the intended and unintended, positive and negative, primary
and secondary effects of a project or program relative to other influences that directly or indirectly
contributed to the impact on rural poverty. The type of questions it seeks to answer, therefore, are:
“what has changed (or not) for whom and why”; “how sustainable are these changes likely to be”;
“what are the impacts and what has caused these changes”; “what has been the program’s
contributions to these changes among other causes”; and “what are the implications for future
program strategy”.
3.
To answer these questions and address the challenge of rigorously assessing and learning about
program contributions to impact from a systemic perspective, PIALA draws on:

a generic impact framework29 that links IFAD’s standard rural poverty indicators to capitals and
relationships and changes in policies and institutions (assessed in a gender & generation
sensitive and disaggregated manner);
27
Under its 9th Replenishment (2012-2015), when the PIALA initiative was launched, IFAD committed to moving 80 million rural
people out of poverty cumulative from 2010 onwards to 2015, and conducting 30 rigorous impact assessments.
28
In the context of IFAD-funded projects and programs, impact implies significant and sustainable changes in rural poor people’s
livelihoods and capabilities to overcome poverty, requiring changes in capacities, relationships, policies and institutions –which is
systemic change. To realize this, IFAD works through governments and in partnership with farmer organizations, civil society,
private sector and international donors. (IFAD, 2011b)
29
PIALA’s generic impact framework was developed and agreed with the IFAD design support of sponsor group in Rome in
October 2012 and was included in the overall PIALA research strategy (IFAD, 2013). It is line with IFAD’s Results- & Impacts
Measurement System (RIMS) and Women’s Empowerment in Agriculture Index (WEAI).
5

a dynamic Theory of Change (ToC) approach that helps visualize the presumed change
pathways, map out program contributions among broader influences on impact, and identify the
assumptions underneath;

multi-stage sampling using PPS that permits inquiry of the effects of different (with/without)
configurations of program mechanisms and their influence on livelihoods and poverty from a
systemic perspective using mixed-methods;

nested mixed-methods for collecting quantitative and qualitative data in relation to the causal
links in the ToC and the evaluation questions around these links, using participatory processes
and triangulation techniques;

a participatory sensemaking model for extensive cross-validation with key stakeholders at local
and national levels, enabling them to critically engage with the evidence and obtaining a broader
systemic view of the processes impacting rural livelihoods;

configuration analysis of heterogeneous patterns of program treatment (or various ‘with/without’
configurations) and their influence on livelihoods to assess program contributions to impact (as
an alternative for counterfactual-based analysis);

a standardized data collation and reporting approach that links the evidence of cascading
changes and causes from impact to program mechanisms using the ToC as a guiding structure.
4.
PIALA is thus a theory-based mixed-methods evaluation approach. ‘Theory-based’ implies looking
at the causal chains and assumptions that underpin the causal flow from program interventions
through intermediate changes to the higher-level outcomes and impacts. The program Theory of
Change (ToC) provides a structure to determine the focus of the impact evaluation, identify the
methods specific to the causal links and questions to be inquired, and engage all stakeholders in
comparing evidence of actual changes and causes with what was envisioned in ‘theory’. The data
needed to answer the evaluation questions are collected through a selection of methods that draw on
each other’s strengths analytically, while compensating for each other’s weaknesses and biases they
show when used separately as a single method or approach.30 Multistage random sampling supports
the nesting of survey-based and participatory methods at grassroots level and in-depth interviews
with market and policy actors at the higher systems levels, as to enable a robust analysis of rural
poverty impact from a systemic perspective. The assumption is that such a nested mix of methods
can produce generalizable conclusions, if applied in a standardized manner across representative
population samples in ways that reduce bias and accommodate heterogeneity.
5.
Participatory methods are often considered though as not suitable for impact evaluation because of
the perceived risk of bias. Participation and debate about evidence of impact to influence policies and
behaviours obviously incite power dynamics. Thus using participatory methods challenges the
independence of the research and thus the reliability of the evidence. To address this challenge,
rigorous facilitation is required of the participatory processes. Acknowledging that an evaluation is
never power-neutral and entirely free from political influence or organizational pressure, and
particularly not when using participatory methods, rigor must be defined broader than in purely
statistical terms and also include quality thinking, sharp observation, power-sensitiveness, engaging
multiple perspectives and systematic cross-checking.31 This is what Chambers (2015) calls “inclusive
30
There is a growing consensus in the international evaluation community that mixed methods is in most cases a better choice for
the design and conduct of impact evaluation than a single quantitative or quantitative method, in particular when it comes to
evaluating more complex programs. (Bamberger, 2012; Stern et al., 2012; Westhorp, 2012; Howard White, 2014; Sarah White,
2015; Chambers, 2015).
31
Cf. IFAD & BMGF, 2013c: 7.
6
rigor”. Using visual tools such as causal flow mapping and matrix scoring is particularly powerful in
this respect, as it enables people to see how the reconstruction of change pathways takes shape and
indicate where things are mostly relevant to them. There is always the danger of one with more
knowledge and power trying to override the others who have less knowledge and power, yet good
facilitators notice this fairly quickly and know how to derail these attempts. They also know how to
arrange the groups and employ the methods in such a way that the voice of the majority becomes
stronger than the one of an individual power-holder.
6.
PIALA also tries to move beyond extractive data collection towards more dialogue by not only
collecting but also analysing and making sense of evidence in more interactive ways. Participatory
sensemaking processes are organised at key levels and units of analysis, before the research at these
levels and in the specific locations is finalised. While almost non-existent in impact evaluation,
experiences in large-scale planning and policy-making suggest that such participatory processes can
help improve relationships between governments, citizens and private actors. Hence PIALA builds
on four key principles:
a) listen to those whose lives are (directly or indirectly) supposed to be improved;
b) avoid viewing participation as simply a mechanisms for better data collection and also involving
different stakeholders meaningfully in the evaluation processes;
c) produce evidence of an project/program’s influences on rural poverty impact that is rigorous,
contested and debated and helps understand the interactions and processes generating (or
hindering) such impact; and
d) amplify the voices of the poorest and less powerful (particularly women and minorities) in the
critical analysis of change processes and an intervention’s influences.
7.
A configuration analysis method is developed for clustering and comparing the large amounts of
qualitative and quantitative evidence obtained with the participatory mixed-methods research. This
method helps examine the patterns of interaction and influence in/between the various change
processes at different levels in the ToC to reach conclusions about program contributions to impact.
This is done by first ‘unzipping’ the ToC to identify the different combinations of program treatment,
outcomes and conditions that is supported by the evidence, and subsequently ‘zipping up’ the
findings again along the ToC to arrive at conclusions. Combining QCA32 techniques with a systemic
ToC approach, this configuration analysis method offers an alternative way to arrive at rigorous
causal inference in the absence of clean control groups. This is particularly useful for
programs/projects where it is quasi-impossible to find such clean control groups, or where
institutional and policy work has purposively ‘contaminated’ all groups.
8.
Finally, in hopes of creating greater value, the PIALA processes and methods were designed and
piloted around three quality dimensions: rigour, inclusiveness and feasibility. Rigour is understood in
terms of methodological consistency and reliability, which in a participatory mixed-methods
approach emanates from both the rigorous employment of methods and the rigorous facilitation of
participatory processes. Inclusiveness refers to the meaningful engagement of stakeholders and the
credibility of findings, requiring rigorous facilitation. Feasibility concerns the budget and capacity
needed to meet the expectations with regards to rigour and inclusiveness. A quality assurance
framework (QAF) was developed and piloted alongside the approach for assessing performance on
32
Qualitative Comparative Analysis (QCA) is a methodology used for analysing large and small n data sets by identifying all
possible combinations of variables observed in the data set, and then applying the rules of logical inference to determine which
conclusions are supported by the data. More information about this methodology can be found on:
http://www.u.arizona.edu/~cragin/fsQCA/index.shtml.
7
these three dimensions in three to four subsequent phases of the evaluation33. The QAF is attached in
Annex 19.
1.2 PIALA training and management
9.
The design of the evaluation and the training of the research team was headed by Adinda Van
Hemelrijck (IFAD consultant) while the management, coordination and field supervision was led by
Glowen Kyei-Mensah (managing director of PDA, the research firm that was commissioned to
conduct the evaluation). The two worked closely together in an equal partnership, both contributing
and taking joint responsibility for the design, quality and results of the evaluation.
10.
Training for the RTIMP impact evaluation took place between October and December 2014, and was
an integral part of the design phase. Blended with design work and field-testing, it involved two
major training phases: a PIALA design training (2 days) and a PIALA methods training with field
testing (8 days).
11.
Although all members of the core research team were well versed in participatory methodologies,
none of them had been involved in impact evaluations combining participatory methods with
statistics. PIALA was therefore new but not entirely unfamiliar. Yet the 2-day design training was
crucial to afford the team leaders with the basics of a PIALA impact evaluation and help them
understand what type of impact assessment was aimed for, what type of questions were to be
answered, the components to be adapted to the context as part of the in-country design, and the
standards and principles to adhere to. It essentially sought to prepare them for a detailed design of the
impact evaluation of RTIMP.
12.
The design training was further extended into a consultative design process. The team spent one
week (5 days) carrying out a thorough desk review and reconstructing RTIMP’s Theory of Change,
and one week (5 days) conducting stakeholder consultations and organising a design workshop,
resulting in the drafting of the evaluation design paper34. After the design workshop, a one-week (or
5-day) PIALA methods training was organised that familiarised the entire research team (including
research assistants) with the existing field methods and tools of PIALA, and introduced them to some
new methods and tools. Furthermore, methods and tools were tested during 3 days in two research
sites in the Eastern and Central Regions and modified before completion and approval by the Client.
13.
The researchers prepared and agreed on translations in the dominant languages before undertaking
focus groups and interviews during field-testing. Based on the reflections on the field tests, tools
were adapted and improved. Improvements were also made in response to the feedback received
earlier from IFAD and external reviewers on the first PIALA pilot in Vietnam. The methods and
tools used in Vietnam were revised and new methods and tools were developed specifically for the
impact evaluation in Ghana. The product coming outcome of this process was the researchers’
handbook that was compiled based on the PIALA guidance developed by the PIALA core design
team.
14.
From this point the research team was fully prepared to undertake the fieldwork. The research team
was divided into three teams, one team per zone (North, Central and South), each consisting of four
researchers speaking multiple local languages. Each team was divided in two sub-teams of two
33
The structure of the QAF was inspired by the “Better Evaluation Rainbow Framework” (cf. http://betterevaluation.org/plan). The
findings from the methodological reflections on the PIALA pilots using the QAF are presented in separate reports.
34
Cf. Van Hemelrijck, A. & G. Kyei-Mensah (2014). Design Paper for the impact evaluation of the Root & Tuber Improvement &
Marketing Program (RTIMP). Participatory Impact Assessment & Learning Approach (PIALA) developed with support of IFAD
and BMGF.
8
working in parallel. The three teams were closely supervised by both the research manager (Glowen
Kyei-Mensah) and IFAD’s evaluation consultant (Adinda van Hemelrijck). Both alternately
accompanied the teams in the North, Central and South during the entire data collection process. The
three teams also reported and sent their field notes to the research manager on a daily basis.
1.3 Cost & benefits of different design options
15.
All impact evaluations come at a serious cost if they are to be done rigorously. Only if
commissioners understand the potential outcomes and benefits of different design options, and are
clear about the type of questions that need to be answered for what purposes, can they make a sound
decision about the level of rigour/quality and thus the budget required. This was particularly
important for an impact evaluation using PIALA, given its demands to sampling for meeting
statistical principles on the one hand, and those of facilitation for participatory data collection and
sensemaking on the other, both of which are resource-intensive in themselves. Combining
participation and statistical analysis in a nested mixed-methods demands high coverage, enough time,
and highly competent researchers, thus is naturally more resource-intensive.
16.
Three design options were therefore presented for discussion to the commissioners before any
ballpark figure was agreed or contract was signed:



The first option was a full-scale but narrow-focused PIALA version that looks at one or two
program mechanisms in the program ToC, but in a country-representative sample. The total cost
of such a design was estimated at US $ 190,000. The purpose is learning about the effects of
one or two particular aspects of the program. Components are studied in isolation, thus not
permitting conclusions about their systemic interactions. Findings would have been insufficient
to report on the program’s total contribution to rural poverty impact, since most parts of the
program would not have been looked at.35 Moreover participants would have learned about
particular aspects of the program but not gained any broader understanding of how their and
other’s actions and interactions in different areas of work affect livelihoods and poverty.
The second option was a full-scope but narrow scaled PIALA version that looks at the full
range of mechanisms in the program ToC but in a limited geographic area. The total cost of
such a design was estimated at US $ 165,000. The purpose is learning about the program’s total
contribution to rural poverty impact under specific conditions. This is a case study approach:
cases are selected on the basis of their learning value. Participants would have learned a great
deal about the systemic interplay of the different components of the program and the actors
involved, but only in these particular cases. Findings would not have been generalizable and
sufficient to report on contributions to rural poverty impact across the entire country or the
entire area covered by the program36.
The third option finally was a full-scope and full-scale PIALA version that takes a systemic
perspective and looks at the full range of mechanisms in the ToC in a country-representative
sample. The total cost of such a design was estimated at US $ 280,000. The purpose is learning
about and reporting on the program’s total contribution to rural poverty impact as a whole,
35
An example is the thematic impact study conducted in 2014 on the FFF-mechanism under RTIMP that recommends “the
expansion or scaling-up of the FFF program across Ghana given the impact the project has had on beneficiary farmers as well as
the requests from non- beneficiary farmers” (MoFA, 2014b: 30). The success of the FFFs however has contributed to market
saturation in the context a downward conjuncture of high inflation, dropping prices and failing infrastructures. This is not looked at
by this study since its focus is merely on the single mechanism FFF and its immediate effects.
36
Unless the program itself has taken a case-based approach, for instance when conducting pilots in different contexts to test new
mechanisms and shape new policies.
9
while zooming in on specific mechanisms with greatest learning value. Here participants would
have learned about the systemic interplay of the different program components across the entire
country, thus expanding their horizon by looking at the program’s influences on livelihoods and
poverty from different vantage points and superseding their own location. The outcomes and
products of such an evaluation can be used for both reporting and learning for/with different
audiences at different levels and for different purposes (e.g. policy making, new program design
or improvement of mechanisms, funding decisions, empowerment and mobilization).
17.
The commissioners of the impact evaluation of the RTIMP –i.e. the MoFA/GoG and the IFAD
Country Office (ICO) in Ghana– jointly choose for the third full-scope and full-scale option for a
total of US $ 233,000. The purpose was to critically reflect on the influence and reach of the various
mechanisms that were employed under the RTIMP and raise issues that merit closer attention and
more innovative thinking under the GASIP. At the same time the evaluation also needed to produce
rigorous quantitative evidence of program’s contribution to rural poverty impact for reporting to the
GoG and IFAD. To further improve and pilot-test the PIALA methodology, an additional
US $ 60,000 was invested by IFAD and the BMGF (of which US $ 20,000 for methodological
innovation and US $ 40,000 for PIALA training, supervision and meta-inquiry).
1.4 Overview of PIALA methods and processes employed in Ghana
18.
The table below presents an overview of the PIALA methods and processes employed in the
evaluation of RTIMP in Ghana, and their purposes (or intended uses) and participants, in three
distinct phases: design, fieldwork and analysis. For each of the three phases, also the total cost and
products are listed. The total net amount of participants without overlap was over 2000 (incl. 837
households, 1180 FGD participants with some overlap with the households in 4 districts, and over
100 KII participants). The various methods and processes and distributions of participants are
described in greater detail in Sections 3-6.
10
Table 1.4: Summary of PIALA processes, methods & tools used in the evaluation of RTIMP
PROCESSES, METHODS & TOOLS
1. DESIGN PHASE:
Focusing and framing
the evaluation
Product:
Evaluation Design Paper37
Total cost:
US $ 17,850
(incl. PIALA design
training and stakeholder
design workshop)
2. FIELDWORK:
Collecting and linking the
data
Products:
 Field Research Manual
with detailed “how to”
guidance for each
method
 District field notes and
data collation tables
Total cost:
US $ 146,000
(incl. PIALA methods
training and field testing)
37
Projection of potential cost-benefits of different
design options (Section 1.3)
Methods/tools:
Outline of 3 design options (full scale–full
scope; limited scale–full scope; full scale–
limited scope) in relation to the 3 PIALA
purposes (reporting, advocacy, learning)
Reconstruction and visualisation of RTIMP’s
Theory of Change (ToC) (Section 2.1)
Methods/tools:
Emerging ToC diagram that shows the
envisioned causal pathways (with codification
of the causal links) elicited from the program
documents and the discussions with national
stakeholders
Sampling and developing the methods and tools
for data collection, data collation and data
quality monitoring (Sections 2.3 & 3.1, Annexes
10-13)
Methods/tools:
 Sampling hierarchy
 Data collection & methods table
 ‘How-to’ guidance for employing the data
collection, collation and quality
monitoring methods
 Standard note-taking formats
PURPOSES
PARTICIPANTS
 Enable commissioners to make a decision
about scale, scope and purpose of the
evaluation based on an adequate
understanding of the different design options
in terms of quality, outcomes and budget
implications
 the IFAD Country Program Manager
 the RTIMP Coordinator representing the
MoFA/GoG
 Identify the program’s impact and
contribution claims to be evaluated, and
formulate evaluation questions focused on
these claims and their assumptions
 Create a shared understanding of the
program’s ToC (including broader influences
on impact)
 Select the methods specifically in relation to
the causal links in the impact and contribution
claims
 National key stakeholders who had been
involved in program implementation,
management and supervision (total of 32
participants –incl. RTIMP Coordination Unit
and Steering Committee, IFAD Country
Program Office and consultants, MoFA, PFIs,
RIs, SCFs)
 Enable a systemic inquiry of the impact of the
combined changes in production, processing
and market linking on livelihoods and poverty
status in 30 random supply chains across the
country
 Enable comparative analysis of the systemic
inquiries of the 30 supply chains
 Ensure rigorous employment of methods and
facilitation of participatory processes
 Ensure systematic data capturing, data
collation, data quality monitoring and
reflective practice during fieldwork
 PDA research team (incl. research assistants),
GSS statistician, 2 methods consultants
Cf. http://www.participatorymethods.org/authors/adinda-van-hemelrijck-and-glowen-kyei-mensah.
11
Data collection on changes in access to food &
income and its causes (Sections 5.2 & 5.3)
Methods/tools:
 Household survey
 Generic change analysis (incl. change
ranking and causal flow mapping of
changes in wealth & wellbeing)
Data collection on changes in R&T livelihoods
and its causes (Section 5.4)
Methods/tools:
 Generic change analysis (see above)
 Livelihood analysis method (incl. change
matrix exercise, causal flow mapping, and
SenseMaker
Data collection on reach and effects of selected
program mechanisms (DSF, FFF, GPC/MEF)
(Sections 5.5 & 5.1)
Methods/tools:
 Livelihood analysis (see above)
 Constituent Feedback (using a specific set
of facilitation and scoring questions for
each mechanism)
 Semi-structured interviews (mirroring the
scoring questions in the Constituent
Feedback)
Data consistency and quality monitoring
(Section 5.6)
Methods/tools:
 Standard data collation table
 Daily team reflections using five standard
sets of questions (use of methods,
facilitation of processes, data capturing,
sufficiency of data on causal links, and
sufficiency of data on program
mechanisms)
 Collect and triangulate data on the link I2I1
in the ToC
discussion of changes in livelihoods affecting
household wealth and wellbeing, based on a
visual reconstruction of the actual causal
pathways
 837 households (random) were surveyed
 439 intended program beneficiaries (quasirandom; 51 % women and 49 % men)
participated in the generic change analysis
 Collect and triangulate data on the link
 400 intended program beneficiaries (quasiO1+O2+O3I2 in the ToC
random; 47 % women and 53 % men)
participated in the livelihood change analysis,
of which 393 did the SenseMaker exercise
discussion of changes in production,
processing and market linking affecting their
(participants in the generic change analysis see
livelihoods, based on the visual reconstruction
above)
of the actual causal pathways
 Collect and triangulate data on the causal
links between the program mechanisms (DSF,
FFF & GPC/MEF) and the observed changes
in production, processing and market linking
(O1, O2 & O3)
 Engage beneficiaries of RTIMP in a group
discussion and anonymous scoring of the
reach and benefits of the services provided
through the program, and the effects of these
on the changes in production, processing and
market linking that affected their livelihoods
 341 beneficiaries participated in the
Constituent Feedback (53 % women, 47 %
men)
 100 district officials and service providers (75
district-level and 25 regional/national)
participated in the KIIs
 Identify data gaps and weaknesses early on
during fieldwork to make it possible for the
researchers to probe for more information in
the sensemaking workshops
 Ensure the evidence is robust according to
standards of mixed-rigor (looking at both
inclusive and scientific rigor)
 Instant data processing and cross-checking
during fieldwork making it possible to
organise debates with local stakeholders
around emerging evidence in district
 PDA field research teams
 supervision by the PDA research coordinator
and the IFAD consultant
(participants in the livelihood change analysis
see above)
12
3. ANALYSIS:
Synthesizing the evidence
and analysing and debating
program contributions
Products:
 District sensemaking
workshop reports
 National sensemaking
workshop report
 Aggregated data
collation table
Total cost:
USD $ 70,000
(incl. workshops
and reporting)
Participatory sensemaking (Section 5.7)
Processes:
3. half-day local sensemaking workshops in
23 of the 25 sampled districts
4. 2-day national sensemaking workshop
Methods:
 reverse engineering
 patches & nodes
 iterative design
 vantage points
 soft systems modelling
Configuration analysis (Section 6)
Methods/tools:
 aggregated data collation table
 configuration analysis method
 Obtain additional information and fill in
remaining data gaps
 Help program stakeholders develop a more
systemic understanding of the development
processes impacting rural poverty
 Engage program stakeholders in discussing
and valuing program contributions to rural
poverty impact, and identifying priority areas
for investment
 Give voice to those who intended to benefit,
while offering decision-makers and serviceproviders the opportunity to engage in
dialogue with these voices, based on evidence
 640 local research participants participated in
the district sensemaking workshops (81 %
intended beneficiaries of which 48 % women
and 52 % men)
 106 local, regional and national participants
(40 % intended beneficiaries of which 38 %
women and 62 % men; 45 % officials; 15 %
private sector actors)
 Arrive at rigorous causal inference in the
absence of clean control groups
 IFAD consultant (PIALA project leader)
 PDA research coordinator and research team
leaders
The participants in the sensemaking workshops
were selected from the field research participants
(with the exception of the households)
13
2
RTIMP evaluation framework
19.
This section describes the design aspects and methodology of the impact evaluation of RTIMP
employing PIALA as its overarching framework. Most of what is described here can also be found
the design paper of the impact evaluation.
2.1 RTIMP Theory of Change (ToC)
20.
A Theory of Change (ToC) lays out the broader picture of how program designers, funders and
implementers envision change to happen, by visualizing the causal pathways of the different program
components, and linking these to the higher-level changes that are expected to generate impact. In
evaluative terms, these causal pathways are called the program’s contribution claims, while the
higher-level changes towards impact are called its impact claim. In the RTIMP ToC, the three main
program components –namely: production, processing, and market-linking reflect the program’s
contribution claims while the goal presents its impact claim. This evaluation was commissioned to
conduct a systemic analysis of the interplay between these contribution claims and the extent to
which these have helped realise the program’s impact claim.
21.
Every program has an implicit or explicit ToC. Also a logical framework is a ToC, but a particular
type of ToC that is more linear in its approach and more focused on the individual performance of a
contracted partner (or of a manager responsible for a specific intervention or program component),
rather than the collective impact of multiple actors and multiple interventions. (Funnell & Rogers,
2011) The type that is used in PIALA takes a more complex systemic approach in its attempt to
visualize the various linkages and feedback-loops between the different program components and
mechanisms and their collective outcomes and impacts.
22.
Such a systemic ToC approach is most effectively used as a dynamic and adaptive framework for
evaluating and managing more complex multi-actor programs, for three important reasons. First, it
allows for a rigorous and systematic assessment of a multiple interacting causal links using
appropriate methods for data collection and analysis that are not necessarily counterfactual-based.
Second, it enables different stakeholders to engage with the evidence collected on these links, probe
their assumptions, and critically analyse and debate their roles and contributions to impact on rural
poverty. Third, if used from the beginning of a program, the approach permits timely corrections or
adaptations based on the learnings from the analyses and debates, while also contributing to the
capacity-building of stakeholders to think and operate more systemically and evaluative, hence
producing greater value-for-money.
23.
Since the RTIMP ToC for this evaluation was developed at the end of the program, instead of at the
beginning of the program’s design stage, naturally it had to be reconstructed based on its existing
theory, and thus reflects the more linear picture of the program components or causal claims as
presented by the program’s logical framework. Yet by visualizing the links between the claims, the
discrepancy between how these components were expected to interact and lead to impact, with how
they actually were implemented (which was rather linear and independent from each other) became
apparent. Although the program’s impact on access to food and income is undeniably, evidence
produced by this evaluation clearly shows that this discrepancy has hampered the program’s ability
to generate greater and more sustainable gains in R&T livelihoods across the entire country.
24.
The RTIMP ToC was reconstructed based on a desk review and a one-day design workshop with
national program stakeholders. In this workshop, participants agreed on the program’s key
14
mechanisms, the assumptions and the questions related to each of the causal claims that the
evaluation would have to focus on. Furthermore, they also indicated where external influences
(positively or negatively) had interfered with the envisioned change processes, and agreed on the
main reference period for the evaluation.38 This section presents the ToC diagram and description.
The decisions regarding the evaluation focus and frame that came out of this workshop are presented
in the subsequent Sections 3.2 and 3.3.
2.1.1 Description of the RTIMP ToC
25.
RTIMP was built on an inclusive value chain development rationale that implied: (a) the
improvement and growth of small R&T production and processing businesses, and (b) the linking of
these local businesses to supply old and new R&T commodity markets. The RTIMP sought to enable
resource-poor farmers and processors to seize new business opportunities emerging from these
markets and develop strong local supply chains that would make Ghana's R&T commodity chains a
strong driver for sustainable and inclusive rural-economic growth.
26.
Through the development of these local supply chains, it was assumed that livelihoods would
improve to the extent that rural poor people living in the R&T catchment areas, and by extension in
entire rural Ghana, would become food and income secure. Hence the program goal was to enhance
income and food security of the rural poor by improving R&T-based livelihoods through building
market-based systems that can generate profitability at all levels of the commodity chains. To realise
this goal, RTIMP focused on enhancing smallholder production, processing and market-linking as
the three main program components.
27.
By gradually commercialising and linking smallholders’ production and processing businesses,
supply chains would be formed that effectively would meet old and new market demands. Access to
improved technologies, certified seeds and standardized equipment was expected to sufficiently
increase production quality and quantity to trigger this change process. Access to business training
and financing and exposure to good practices would enable smallholders to develop profitable
businesses and accelerate the growth of smallholder economies at scale. To trigger and enable these
change processes, the RTIMP employed a number of funds and mechanisms, some of which will be
taken forward and scaled up in the government’s next GASIP in different shapes and formats,
including:

District Stakeholder Forums (DSF) for addressing the supply and demand issues as well as
and the technical problems of supply chain actors, and link them to old and new markets;

Supply Chain Facilitation (SCF) for helping develop the supply chains and link them to larger
and new markets with the aid of a small initiative fund;

Farmer Field Forums (FFF) for involving resource-poor farmers and seed growers in fieldtesting and demonstration of improved seeds and technologies and developing a basis for farmer
organization and commercialization;

Micro-Enterprise Fund (MEF) for co-financing of resource-poor supply chain farmers and
processors through the establishment of a matching grant; Information, Education and
Communication (IEC) for informing intended beneficiaries about R&T commodity chain
support services and engage them in program activities;
38
The workshop was organised on 12th of November 2014 in Kumasi. Around 40 people from the various program stakeholders
participated, including: RTIMP coordinators and Steering Committee; IFAD country program manager, senior staff and
consultants; MoFA national and regional directors and officers; PFI representatives, researchers, the SCFs and TREND.
15

Good Practice Centres (GPC) for demonstrating and promoting good processing, quality
management and business development practices, using improved technologies and standardized
equipment.
2.1.2 Visualisation of the RTIMP ToC
28.
The diagram presented on the next page visualizes the RTIMP Theory of Change as described above.
All the boxes in this diagram were coded so that appropriate methods and processes could be
identified for assessing different types of links between different types of changes/causes.
29.
The numbers in the codes (1-3) correspond with the program components in the logical framework
which in the diagram reflect the program’s contribution claims, namely: enhanced market-linking
(component or contribution claim 1); enhanced production (component or contribution claim 2); and
enhanced processing (component or contribution claim 3). The impact claim is formed by the
outcomes of these three contribution claims and the impacts envisioned in the program goal
(O1+O2+O3→I2→I1). The letters reflect the type of change/cause, namely:





30.
‘I’ depicts the two levels of impact presented by the program goal;
‘O’ stands for the higher-level outcomes that together would lead to the impacts;
‘C’ are the expected changes or effects of the program mechanisms;
‘M’ are the mechanisms put in place by the program to generate the expected changes;
‘E’ are the external influences identified by the design workshop participants (the codes in
subscript correspond with the changes/causes in the ToC diagram affected by these influences).
More influences have been identified during the evaluation than in the design workshop and
indicated in the diagram above. These will be described alongside the evaluation findings in Sections
7, 8, 9 and 10. Those indicated in the ToC diagram are the following:
EO1. Provision of infrastructure in the form of feeder roads by the GOG and District Assembly.
EO2. Policy inconsistency related to free seed distribution, hampering commercialisation.
RTIMP distributed certified seeds freely in its early start-up years, until 2010 when a Program MidTerm Review was conducted that recommended a commercialisation of certified seed multiplication
and distribution supported by the Good Practice Centres. Between 2010 and 2012 a transitioning took
place, in which free seed distribution coincided with commercialised production and sales. Only from
2012 free distribution by RTIMP has stopped. Other programs such as WAAP have continued
distributing free planting materials.
EC1. Shift in policy and practice from subsistence to commercial farming threatened by CC.
Prior to RTIMP, cassava was considered as merely a food sufficiency crop. By focusing on increased
production and market-linking, RTIMP has triggered a shift in policy and discourse that moved away
from free hand-outs and government subsidies and positively influenced the commercialisation of
R&T production in Ghana. Climate change however is threatening this process and likely has
negatively affected its outcomes.
EC3b. Influence of the IFAD-funded Rural Enterprise Programme (REP).
REP has built the capacities of the Participating Financial Institutions (PFIs) and their local branches
(e.g. in loan appraisal and disbursement techniques), which should have positively contributed on the
outcomes of Micro Enterprise Fund (MEF) of RTIMP.
EC3c. Lack of regulatory procedures and institutions needed to ensure proper regulation and quality
enhancement of R&T production and processing to meet new market standards.
16
Figure 2.1.2: RTIMP Theory of Change diagram developed and used for this impact evaluation
M2c: Farmer Field Forums (FFF) engage
farmers, extension agents and researchers in
developing, demonstrating and promoting
appropriate R&T production technologies
C2b: Resource-poor R&T farmers organise
and register as FBOs that can access credit
and bargain better market prices
M2b: Training & starter pack for commercial
seed growers to multiply certified R&T seeds
C2a: Resource-poor R&T farmers & seed
producers gain access to and adopt improved
R&T seed varieties, technologies & inputs to
improve crop husbandry, soil fertility and
pest management practices
M2a: R&D for developing bio-agents
EC1
M1c: Information, Education &
Communication (IEC) about CC support
services, inputs and technologies
C1b: Resource-poor R&T processors, farmers
& seed producers commercialize and establish
effective supply chain linkages
M1b: Supply Chain Facilitation (SCF) and
market linking through the Initiative Fund (IF)
EC1
C1a: R&T supply chain farmers & processors
are capable of developing and implementing
viable business and marketing plans
M1a: Training of resource-poor farmers and
processors involved in the R&T supply chains in
business development and marketing
M3c: Co-financing of R&T supply chain farmers
and processors by matching 40% RTIMP funds
with 50% loans from PFIs and 10% self-financing
through the Micro-Enterprise Fund (MEF)
EC3c
M3b: Subsidized upgrading of advanced R&T
processors into Good Practice Centres (GPCs)
that demonstrate and promote good quality
processing & management practices
M3a: Training of artisans to produce and maintain
standardized processing equipment
for R&T supply chain processors and GPCs
EO2
O2: Enhanced
R&T productivity
and production
at scale
EC3b
EO1
O1: R&T supply chain
actors effectively solve
their supply & demand
issues and timely obtain
technical support,
resulting in sustainable
and inclusive CCs
linked to old and new
markets
I1: Rural poor people
in CC catchment areas
have increased access
to food & income to
sustain an active and
healthy life
I2: Improved R&Tbased livelihoods for the
rural poor in CC
catchment areas
C3c: R&T supply chain farmers and
processors gain access to business financing
and market-linking services
C3b: R&T supply chain processors gain
access to and adopt standardized processing
technology and good quality
management practices
C3a: R&T processors grow and develop into
GPCs that are profitable enterprises
O3: Enhanced R&T
processed volumes of
high quality at scale
M1: District Stakeholder
Forums (DSFs) for addressing
supply & demand issues and
technical support needs of R&T
supply chain actors members
17
2.2 Evaluation focus
31.
The impact envisioned by RTIMP is reflected in its goal statement, namely: “enhanced income and
food security of rural poor households through improvements in R&T-based livelihoods and
strengthened market-based systems generating profitability at all levels of the commodity chains”. At
the design workshop, it was proposed to redefine “enhanced income and food security” to avoid a
too narrow interpretation of food security as ‘food self-sufficiency’ and ensure sufficient attention
would be paid to the profitability and sustainability aspects of improvements in R&T livelihoods and
market systems. Hence impact was defined in terms of “access to food and income to lead and
sustain an active and healthy life”, thus focusing impact-level data collection on essential changes in
food, assets, income, R&T revenues and R&T activity.
32.
While the production component was started much earlier in the Roots and Tuber Improvement
Program (RTIP)39 that preceded the RTIMP, interventions related to enterprise upgrading and
market-linking were added under RTIMP, some of which became effective on a national scale only
after the 2010 Mid-Term Review (MTR). Hence it was agreed in design workshop to focus the
evaluation mainly on the last 5 years, starting from the MTR (2010-2015) for evaluating the interplay
between the three components. The 2008 RIMS baseline was used for comparison of findings merely
related to enhanced R&T production.
33.
Furthermore, it was also agreed to focus the evaluation on the four main types of commodity chains
that had been developed in this reference period, namely: Gari, High Quality Cassava Flour (HQCF),
Plywood Cassava Flour (PCF) and Fresh Yam for Export (FYE).
34.
Lastly, it was agreed to focus on the four main program mechanisms that would be considered for
scaling up in the new GASIP –namely: the District Stakeholder Forum (DSF), the Farmer Field
Forum (FFF), the Good Practice Centre (GPC) and the Micro-Enterprise Fund (MEF).
2.3 Assumptions and evaluation questions for each causal claim
35.
The matrix below outlines the main evaluation & learning questions relative to the assumptions
underneath each of the causal claims in the RTIMP ToC. These questions were selected and agreed
by the participants in the design workshop to guide and focus the impact evaluation. Furthermore the
matrix also identifies the methods that are used for data collection at each level of inquiry relative to
each of the links in the causal claims. The selection and nesting of methods is described in greater
detail in Section 1.1.
36.
The causal links are listed in the left column of the matrix (the codes in the links correspond with the
codes in the boxes in the ToC diagram). The middle columns present the methods and processes for
each level of inquiry (including household level, community cluster level, district level, and
zonal/national level). The right column summarizes the sampling approach for each claim. The
sampling strategy is described in greater detail in Section 1.1.
39
The RTIP focused primarily on cassava research and development. The RTIMP extended this focus to other roots and tubers and
added a strong marketing component designed to improve poor farmers' access to food and income.
18
Impact Claim - Poverty reduction
Hypothesis:
Enhanced production (O2) + enhanced processing (O3) + sustainable and inclusive CC supply chain linking (O1)
=> improvement R&T livelihoods (I2) and poverty reduction (I1) in rural Ghana
Assumption:
 Livelihoods and poverty status in rural Ghana can be improved by commercializing smallholder R&T production and processing
businesses combined with the establishment of competitive market-driven and inclusive CC linkages.
I1
O2
I2
O1
Evaluation/learning questions:
O3
 To what extent and for whom does this assumption hold true (or not) and under which conditions? Does it hold true for resourcepoor women and youth/young adults in remote rural areas? What conditions need to be changed to enable women and young adults overcome cultural and socioeconomic barriers?
Evaluation focus: livelihoods improvements and secured access to food and income, particularly for women and youth/young adults.
Causal link
I2→I1
House-hold level
Community cluster level
District level
HH survey
(with households in the
the supply chain
catchment area)
O3+O2+O1→I2
Generic change analysis
(in gender/age-specific focus
groups of community members from
the supply chain catchment area)
40
Zonal & national level
Sampling approach
Review of the 2010 Ghana
Living Standard Survey
report and other relevant
secondary data40
Stratified sampling of 30
households from the
community clusters in each
sampled district
Review of RTIMP RIMS
baseline ad other M&E data
Stratified sampling of
community members from
the community clusters in
each sampled district
E.g. from the Ghana Statistical Service, Food and Crops Research Institutes (CSIR), FAO, WB, UNDP, etc.
19
Contribution Claim of Component 1 – Market linking
M1c
Hypothesis:
Promotion and facilitation of R&T supply chain linking of farmers and processors (M1c+M1b->C1b and M1)
+ business capacity-building of farmers and processors (C1a) => sustainable market-driven & inclusive CC linking (O1)
M1b
C1b
M1a
C1a
O1
Assumptions:
 Sustainable and inclusive CC linkages can be established by building business and marketing capacities among supply chain
processors and farmers and creating a platform or District Stakeholder Forum (DSF) where they can discuss these needs and demands.
 More recourse-poor farmers and processors (incl. women and youth/young adults) will participate in the DSFs and sign up for market-linking
support services if they sufficiently increase their production quantity and quality and are sensitized about the benefits and opportunities.
M1
Evaluation/learning question:
 To what extent do these assumptions hold true (or not)?
 What enables or towards DSFs to become viable “chambers of commerce” –i.e. member networks that serve as private business linking and marketinformation platform empowering buyers, producers and processors (incl. women and young adults) to address their demand & supply issues independently?
 What are the main barriers to linking resource-poor farmers and processors (incl. women and young adults) to old and new R&T commodity markets? What
conditions need to be in place to help them overcome these barriers? What is missing (e.g. certification, packaging, traceability, market prospection)?
Evaluation focus: linkages with old & new markets; CC inclusiveness; reach & benefits of participation in DSFs.
Causal link
C1a+(M1)→O1
C1b+M1a→C1a
House-hold level
Review of DDA reports
Constituent feedback
(with mixed groups of (non-)DSF
participants)
M1c+M1b+O2
+O3→C1b
Livelihood analysis
groups with supply chain farmers
and processors)
District level
KIIs with DDAs, BACs and
supply chain leaders (SMEs,
GPCs, aggregators and
exporters)
KIIs with Supply Chain
Facilitators (SCF) and the
off-takers (industries, food
traders…)
Review of RTIMP Enterprise
Record Books (ERBs), ZOCs
progress reports; MoFA and
DADU Organisational
Capacity Assessments;
RTIMP M&E data (incl.
2014 thematic impact studies
on DSF & SCF and IEC)
Sampling approach
Proportional sampling
of 25 districts in the
catchment areas of the 4
types of commodity
chain across the 3 main
agro-eco zones
Identification of max. 30
community clusters in
the 20 sampled districts
supply chain farmers,
seed growers and
processors
20
Contribution Claim of Component 2 –Enhanced R&T production
Hypothesis:
Communication and participatory R&D on new technologies (M2c+M1c) and production & distribution of certified seeds
and bio-agents (M2a+M2b) => enhancement and scaling of smallholder R&T production (O2)
Assumption:
 Resource poor R&T farmers and seed producers can become commercial growers by organising into FBOs
and adopting improved production technologies. FFFs encourage them to do so.
M2c
C2b
O2
M2b
C2a
M2a
M1c
Evaluation/learning questions:
 To what extent and for whom does this assumption hold true (or not) and under which conditions?
 Do FFFs sufficiently reach more vulnerable and/or illiterate resource-poor farmers (incl. women and young adults) and help them overcome barriers to
participate? What motivates41 resource-poor farmers and seed producers (particularly women and young adults) to participate in the FFFs?
Evaluation focus: FFF reach and farmers’ & see growers’ commercialisation.
Causal link
C2a+C2b→O2
Household level
Livelihood analysis
and processors)
M2a+M2b+(M2c)
+M1c→C2a
M2c→C2b
KIIs with FFF facilitators,
extension agents, District
Development Advisors
(DDAs), and officers from
the District Agricultural
Development Unit (DADU)
Constituent feedback
District level
Review of RTIMP
productivity surveys and
progress reports from the
SRID, GLDB, DDAs and
ZOCs
KIIs with research team
leaders42 of the regional
research institutes
(CSIR/KNUST/UCC)
Sampling approach
seed growers and
processors
FFF- and non-FFFparticipants
Review of RTIMP M&E data,
including the 2014 thematic
impact assessment of FFFs
(with mixed groups of (non-)FFF
participants)
41
42
This concerns an important condition that was mentioned in the consultative design workshop, which needs to be assessed as part of the plausible explanation of program contributions to O2.
There are 7 research team leaders, involved in the FFFs, 5 of which are based in Kumasi, 1 in Cape Coast, and 1 in Tamale.
21
Contribution Claim of Component 3 – Enhanced R&T processing
Hypothesis:
Access to business financing and market-linking services (M3c+M3b->C3c) + exposure to good practices (C3c+M3b->C3b)
=> development of profitable processing enterprises by R&T supply chain farmers and processors
=> enhancement and scaling of smallholder R&T processing (O3)
M3c
C3c
M3b
C3b
M3a
C3a
O3
Assumptions:
 Resource-poor processors and farmers who are well trained in quality management, business planning and marketing apply for matching grant funding (MEF) to
invest in their businesses. PFIs are prepared to provide credit to well-trained resource-poor processors and farmers up to 50% of their planned investments.
 GPCs can reach and teach resource-poor processors to develop more profitable agri-processing businesses by demonstrating good quality processing and
management practices, including the use of improved technologies and standardized equipment. As a result, resource-poor processors apply to the MEF and
invest in new technologies and equipment that help them to produce greater volumes of higher quality at lower cost.
Evaluation/learning question:
 To what extent and for whom do these assumptions hold true (or not)?
 What conditions need to be in place for GPCs to become profitable and attractive businesses particularly for young adults living in remote areas? What supports
or hinders GPCs to better link the supply chain farmers to old and new markets, and how is this influenced by the DSF?
 Reach and added value of the MEF? Effects of the MEF on growth of the funded agro-processing businesses? Avoidance of elite-capture?
Evaluation focus: GPC’s and MEF’s reach and contribution to market-linking; processors’ loan-taking and commercialisation.
Causal link
M3b→
C3a+C3b→O3
Household level
Livelihood analysis
and processors)
District level
KIIs with GPCs
Review of RTIMP and REP
M&E data and supervision
reports (incl. the 2014
thematic impact studies on
MEF and GPC); the
comparative case study on
matching grant facilities
Sampling approach
seed growers and
processors
GPC- and non-GPCparticipants (incl. MEF
beneficiaries)
22
3
Sampling and community mobilization
3.1 Sampling approach
3.1.1 Principle unit of analysis and sample
37.
As already explained in Section 2.2, this evaluation was expected to conduct a systemic analysis of
the interplay between the production, processing and market linking components of RTIMP and its
impact on livelihoods and poverty status for the 4 main types of commodity chains that the program
has developed in the period between 2009/10 and 2014/15 (Gari, HQCF, PCF and fresh yam for
export). The principle unit of analysis and thus sampling population for this were the catchment or
‘supply chain’ areas of the commodity chains. Supply chains consist of ‘supply chain leaders’ (such
as gari and HQCF producing GPCs and factories, plywood factories and fresh yam exporters) and
‘suppliers’ (smallholder producers and processors), and are geographically defined by their location
(in particular of the suppliers).
38.
Since the supply chains were administratively served at the district level, 25 districts were randomly
sampled from the 67 treated by RTIMP43 at the time of the evaluation design. The districts were
sampled across 3 agro-ecological and administrative zones (including the North which is savannah,
the Centre which is transitional, and the South of Ghana which is deciduous forest). The 25 districts
comprised 30 community clusters, each cluster comprising 3 communities and locating a supply
chains.44 The 30 community clusters contained samples of supply chains of the 4 commodities with
probability proportional to seize (PPS) of their total populations of supply chains. Also sufficient
coverage of heterogeneity in program treatment was obtained by ensuring that all possible
with/without configurations of the evaluated program mechanisms in different gradations were
included in the sample. Table 3.1.1 presents the original sample that was taken from the list with
supply chain leaders obtained from RTIMP. The sample frame is presented in Annex 5.
Table 3.1.1. Sampled districts and community clusters
Zone
North
Region
Northern
Upper West
Volta
Centre Ashanti
Brong
Ahafo
Eastern
District
Central Gonja
East Gonja
Nanumba North
West Gonja
Wa East
Nkwanta South
Adansi South
Ashanti Mampong
Kumasi Metropolitan Assembly
Ahafo Ano South
Kintampo South
Pru
Tano North
Tano South
Techiman
Birim Central
Suhum Craboa Coalta
Cluster community
Yapei
Sisipe
Bimbila
Damango
Gulemga
Krotang
Okyerekrom
Mampong
Kumasi
Adesewa
Alora
Zabrama
Nkwanta South
Apesika
Techiman
Otaipro
Amanase
Supply chains
PCF
FY
FY
Gari
Gari
Gari
Gari
Gari + HQCF
PCF
Gari
FY
FY
HQCF
Gari
2 Gari + 1 FCF
Gari
Gari
# clusters
1
1
1
2
1
1
1
2
2
1
1
1
1
1
3
1
1
43
The programme’s completion report says that the programme had worked in 106 districts across all ten regions by the end of its
operational period. At the time of the evaluation design though a list of 68 treated distrocts was provided by the program
coordination unit for sampling.
44
Some deviations (discussed in Section 3.3.1) occurred in the supply chain samples though, which reduced the amount of
researched supply chains from 30 to 25 (corresponding the 25 districts).
23
South
Central
Volta
Western
Total
Upper West Akim
Adeiso
Abura Asebu Kwaman Kese
Abakrampa
Agona East
Mankrong
Assin South
Assin Dominase
Gomoa East
Gomoa Obuasi
Ho Municipal
Ho
North Dayi / Kpando
Wusuta Anfoega
Wassa Amenfi East
Samreboi
17 Gari, 5 PCF, 4 FY, 4 HQCF
Gari
HQCF
Gari
Gari
Gari
HQCF
Gari
PCF
1
1
1
1
1
1
1
1
30
3.1.2 With/without configurations as ‘control’
39.
With RTIMP effects spilling over and many other rural livelihood programs influencing rural
people’s lives and livelihoods all over Ghana, it was very difficult (if not impossible) to find ‘clean’
communities and households that could serve as control groups for determining the net attributable
impacts of RTIMP on household poverty. By ‘clean’ we mean ‘not treated, confounded or influenced’
by RTIMP or any other program that works to reduce poverty through strengthening R&T
livelihoods. Without such a ‘clean’ control group, it is not possible to assess the net difference in
impact. There was also no interest among the core-learning group (established with sponsors and key
stakeholders at the design workshop in Kumasi on 12 October 2014) to collect evidence on
household-level control groups at the cost of a systemic inquiry of the four populations of supply
chains. Hence it was decided not to conduct a classic counterfactual inquiry of rural poverty impact
at the household level, but instead to conduct a configurational analysis of the effects of different
patterns of program treatment (or different ‘with/without’ configurations of program mechanisms) on
changes in R&T livelihoods that impacted household food and income. The evaluation was framed as
a learning exercise and thus sought to understand the explanations for their contributions in terms of
reach, effectiveness and sustainability.
40.
Thus the focus of inquiry was on the relative contributions of the selected program mechanisms to
the R&T livelihood changes that impacted on rural poverty. The sample also included several
districts where the mechanisms were mostly dysfunctional or not in place, which formed a useful
comparison group that provided counterfactual evidence at the level of the observed R&T livelihood
changes (but not at the household level). The evaluation was framed as a learning exercise and thus
sought to understand the explanations for their contributions in terms of reach, effectiveness and
sustainability.
3.1.3 Sub-samples of households and research participants
41.
To inquire trends in household food and income and the influences of changes in R&T livelihoods on
these, a brief household survey was conducted. For this, 30 households were randomly sampled in
each of the 30 sampled community clusters, by systematically selecting every 5th or 10th household
with at least two members, using as a rule of thumb a 60/40 ration of primary to secondary
beneficiaries. Primary beneficiaries were those that the program directly intended to reach and
benefit, which included all households that had R&T as a most important livelihood activity and had
been or still were resource-poor. Secondary beneficiaries were all other households living in the
communities of the R&T catchment areas, since the assumption of the program was that changes in
R&T based livelihoods would also indirectly impact all other poor households living in these
communities.
42.
The households were selected by first canvasing the community and paying particular attention to the
layout of the houses and the different neighbourhoods in the community, taking particular note of the
areas were the rural poor lived. This informed the teams’ decision on choosing the main axes or
24
directions to walk starting from the centre (often the main access point to the road, or the central
school or church) in order to count and select every 5th or 10th household. Once the starting point and
the axes were identified, then the rule was to move in a systematic manner and be consistent with
household selection in all communities.
43.
Last, within each of the 30 sampled community clusters, an average of 42 intended beneficiaries
were selected to participate in the participatory research, using an 80/20 ratio of primary to secondary
beneficiaries and a 50/50 gender ratio with 10-20% young adults (<35 years). Where possible, these
participants were selected quasi-randomly from a list of beneficiaries obtained from the district
officials, or alternatively if no lists were available, by using the snowballing technique (cf.
community mobilisation approach in the next section).
3.2 Community mobilization approach
44.
Generally, the teams were able to mobilize the participants quite independently, without interference
of RTIMP. Mobilisation was done by asking district officials, GPC leaders, and local leaders for lists
of potential beneficiaries of RTIMP, and through community entry visits. Before fieldwork began,
the research team was given the list of RTIMP district officials from the sampled districts. This list
was used to contact the respective DDAs and desk officers in specific districts. First contact with the
focal communities was used to select the two others communities that were part of the supply chain
to make up each community cluster. Where needed, researchers were aided by the officers.
45.
District RTIMP officials assisted the mobilization process by providing useful details of the program
in the districts as well as a comprehensive list of beneficiaries. KIIs were held with district officials
at the beginning of the fieldwork in every district, in order to get a good understanding of the context
and obtain sufficient knowledge about the various communities and the activities that had been
implemented by RTIMP as well as other programs. The list of beneficiaries then formed a starting
point for the research team to select the participants for the FGDs. Where this list was available,
research participants were quasi randomly stratified from the list. In districts where a comprehensive
list was not available, the snowballing technique was used (selected beneficiaries are giving names of
other beneficiaries who on their turn are giving more names of other beneficiaries, etc.).
46.
Prior to study initiation, the research team visited the communities in each cluster. Participatory
techniques were employed, beginning with community entry where the local traditional leader of the
community was the first point of call. The leaders (chiefs, elders or local authority) were explained
that the team was there to carry out a research on changes in R&T livelihoods that had occurred in
their community over the last 5 years. However, to avoid political influence and pressure, it was not
revealed that the research concerned an evaluation of RTIMP at this early stage. Through the process
of community entry, formal consent was obtained to start data collection. The teams then solicited
the assistance of community members to sort out logistics such as the selection of an accessible
central meeting space that is neutral and safe for organising the focus groups discussions and a venue
or place suitable to hold the sensemaking workshop. Focus group discussions took place in a quiet
neutral location at a convenient time for respondents (ensuring their livelihood activities were not
unduly disturbed). Next, using information gleaned from the district RTIMP officers and
beneficiaries, key informants were identified, approached and interviewed.
25
3.3 Challenges in sampling and mobilization
3.3.1 Deviations in the sampling of the supply chains
47.
In some cases, the supply chains turned out to be different in reality from what was sampled on paper,
and/or suppliers for different supply chain leaders could not be discerned, creating deviations from
the original sampling. Wa East was originally sampled as a Gari cluster, but turned out to
predominantly supplying cassava flour for local consumption, and therefore was classified as “other”,
thus not belonging to one of the four main commodity chains. In Techiman, all three 3 sampled
community clusters were producing Gari and suppliers for different supply chain leaders could not be
discerned, therefore researched as one single Gari supply cluster. Tano North appeared to have also
Gari alongside HQCF and thus was researched as two supply chains instead of one. The Gari
suppliers in Tano South could not be discerned from the ones in Tano North and thus were
researched as one single Gari cluster in Tano North. Similarly, the two Gari chains in Damango were
researched as one. HQCF and Gari in Ashanti Mampong were produced by the same GPC, but the
suppliers or intended beneficiaries spoke only of Gari when researching the effects on their
livelihoods, hence the two clusters were merged and researched as Gari.
48.
In the two PCF clusters in Kumasi Metropolitan Assembly, as well as the PCF chain in Wassa
Amenfi West no suppliers could be identified or located; only off-takers or industries were found
without any linkage to the cassava farmers and chips suppliers. In these two districts, data was
collected as much as possible in the way it was done in the other districts, but no district sensemaking
workshops were organised since there was too little for participants to discuss.
49.
To conclude:
 From the five sampled PCF supply chains (plywood), only two were real supply chain areas
(Central Gonja and Wassa Amenfi West), while two were merely industrial off-takers in the
same area (Kumasi Metropolitan Assembly), and one was Gari (Techiman).
 From the four sampled HQCF supply chains (high quality cassava flour), one turned out to be
gari (Mampong), and only three were HQCF (Abura Asebu, Tana North and Ho).
 The sixteen sampled Gari supply chains ended up to be fifteen researched Gari clusters.
 As a result of these deviations, the originally 30 sampled supply chain areas were reduced to 25.
3.3.2 Deviations in the sampling of households
50.
Although the supply chains were reduced from 30 to 25, the original sample of 30 community
clusters was upheld for subsampling the households, as to ensure the total sample size would be
sufficient to arrive at 95 % statistical precision. The intention was to conduct 900 household surveys
within the sampled supply chains that would allow for a multi-variant analysis. In the 2 clusters in
the Kumasi Metropolitan Assembly, however, no suppliers could be found and thus no households
sampled as ‘intended beneficiaries’. Moreover, three household surveys could not be accounted for,
which brouht the total amount of surveys down from 900 to 837 (184 in the Northern, 424 in the
Central and 229 in the Southern zone). This however hasn’t affected the statistical precision.
51.
One of the major challenges encountered was that the vast majority of towns and villages in Ghana
are not laid out in an orderly fashion. The settlements are mostly scattered and houses are not built in
a structured layout. In some cases, it was almost impossible to follow a straight line or direction and
count every 5th or 10th house in any of such direction. Some settlements in the North were widely
scattered and there were few houses, which made it difficult to use the 10th /5th rule. Exceptionally, in
case of very few houses, every 3rd house was randomly selected. Also, in more urban areas there
26
were shops, sheds and other structures in between houses, which made the exact counting more
challenging. Generally the teams stuck to the rule but in some communities they had to use their
discretion while consulting their supervisors. In the Northern Zone, the 10th house method was used
almost everywhere, but this was not so for the other two zones. In some areas (for example Assin
Dominase), the zigzag method was employed to determine the boundaries of the community using a
signpost.
52.
In a few cases there was an overlap between the focus group participants and the household survey
respondents. This happened in the first round of districts that were researched, but was afterwards
corrected. In some places, it was difficult to find households with intended beneficiaries. Where these
could not be found, the research team had to resort to focus group participants. Also at the start of
field work there appeared a misunderstanding among some of the researchers about the concept of
‘intended beneficiaries’, which was interpreted as ‘effectively reached beneficiaries’, which made it
more difficult to find these and caused a selection bias. All households that had R&T as a most
important livelihood activity and had been or still were resource-poor in principle had to be
considered as ‘intended beneficiaries’, no matter if they had been reached by the program or not.
3.3.3 Deviations in community mobilization and the sampling of research
participants
53.
Mobilisation and transportation of research participants to the central locations where the FGDs were
organised (mostly at the GPCs or central market hub locations) appeared quite challenging. A major
constraint was the limited time and budget for mobilisation. Compensation for transportation was
budgeted at 4 GHS per person, but many communities were very remote and public transportation
turned out to be much more expensive. Hence the teams often had to use their own vehicle and
scheduled the FGDs in such a way that the participants in the different communities could be picked
up and brought back in time by their driver. This was quite challenging as 4-6 FGDs were conducted
per day with participants from different communities, and the teams only had one vehicle each. In
three instances, the team vehicles broke down and had to be repaired, which caused substantial
delays in the research schedule.
54.
Moreover there was no budget for allowances to compensate for participants’ time. To avoid creating
the wrong incentive for participating in the research and thus generating biases, the principle of
voluntary and non-paid participation was strictly applied in this impact evaluation. The researchers
experienced great difficulty though in getting this principle understood and accepted by the
participants, as most of them who had been involved in RTIMP were used to receive generous
allowances. Generally, people were not very keen on participating in something that would not give
them a direct and tangible profit or benefit. In many communities, people also felt reluctant to
participate because RTIMP had promised them markets if they would participate in the program, but
these markets mostly didn’t come and thus naturally people felt discouraged to participate in the
research.
55.
In some districts, mobilisation was particularly challenging due to the resistance of district officials,
distrust among the different stakeholders, bad roads making it quasi impossible to convene farmers
and processors from different communities in the supply chain, and the lack of a convening location
where people would feel safe and confortable to talk freely. In Agona East District (Central Region),
North Dayi/Kpando (Volta Region), Suhum (Eastern Region) and Tano North (Brong Ahafo
Region), for instance, the GPC was not functioning and/or there were tensions or mistrust between
the people and the GPC, which made it impossible to use the GPC as a central location for the FGDs
and the sensemaking workshop. In Wasa Amenfi East (Western Region), there was no RTIMP
27
activity, only an industrial off-taker, so there was no reference point to identify intended beneficiaries
and convene them at a hub location that is familiar and connects the participants to each other. In
Mangpong (Ashanti Region), farmers were restrained to participate because of an extension officer
who initially refused to support the research and openly stated that “he owned the farmers”. In Pru
and Kintampo South (both in Brong Ahafo Region), it was very difficult to mobilize the participants
due to bad roads.
56.
Also funerals, market days and farming days made it difficult to consecutively schedule the field
research from one district to the other. Participants didn’t show up at FGDs if there was an important
funeral to attend, they had to bring their produce to the market, or if they had to go to their farm to
work the land the entire day, all of which took place mostly at relatively far distances from their
homes. Hence fieldwork was sometimes delayed during the weekends when there were funerals, and
during the workweek when there were farming and market days. This caused cumulative delays in
the entire evaluation schedule. The teams were able to limit the delays by creatively planning and
sometimes rescheduling the days in a particular district, and even reshuffling the sequence of the
methods whenever needed and possible. This could have affected consistency in the collation and
interpretation of the data, if the different types of methods weren’t selected so squarely to assess
specific links at the different levels of causality in the ToC, and if the questions weren’t so clearly
related to the links.
57.
Yet all these transportation, participation and planning problems did affect the ability of the
researchers to subsample and mobilize sufficient participants within the short timeframe they were in
a district. Also the participatory research participants were subsampled in the original sample of 30
community clusters in the 25 sampled districts, minus the 2 clusters in the Kumasi Metropolitan
Assembly where no ‘intended beneficiaries’ could be found. But focus groups sometimes had fewer
numbers of participants than anticipated due to the problems described above. In a few situations of
the Livelihood Analysis method, the FGDs had to be carried out with mixed groups of both men and
women, as there were insufficient participants to form two separate gender-specific groups. In these
situations, extra attention was paid to the process to ensure all necessary protocols were observed,
women and men had equal opportunities to speak, and responses of both genders were equally
captured in a disaggregating manner. Time periods for these FGDs had to be extended because of
these peculiarities.
58.
Also for the constituent feedback method, which required mixed groups of average 10 participants
(half women and half men, beneficiaries of the DSFs, FFFs, GPCs and MEFs), often not enough
participants could be found in time. In some districts, constituent feedback for GPCs, FFFs or DSFs
could not be carried out because the mechanisms were dysfunctional and thus no participants were
available. In other cases, the tight research schedule conflicted with the processors’ schedules,
especially of those in individually owned establishments or those who had been removed from the
supply chain but were still residing within the supply chain area. Despite all these issues, the
researchers still managed to collect sufficient and useful data from 43 mixed FGDs with a total of
341 participants (thus average 8 per group; 53 % women and 47 % men) complementary to the data
from the other methods, in order to demonstrate the value of the methodology.
28
4
Background and distribution of participants and respondents
4.1 Poverty status
4.1.1 Participatory characteristics of ‘wealth & wellbeing’
59.
Wealth & wellbeing ranking is a PRA method used to collect and analyse perceptual data on wealth
and wellbeing in a community as the basis for identifying locally relevant indicators. Since it was
introduced in the 1980s, this method has become an increasingly accepted means for identifying
locally relevant indicators of wealth and wellbeing and analysing changes in relative poverty status.
The assumption is that by using locally defined indicators of wealth and wellbeing, this helps
overcome researcher biases that may influence the outcomes of the research.
60.
In this evaluation of RTIMP, a wealth and wellbeing ranking exercise was conducted as part of the
participatory generic change analysis method45 that was employed in 23 districts in separate female
and male FGDs. Its purpose though was not so much to collect perceptual data on wealth and
wellbeing for aggregated analysis, but rather to help participants create a shared understanding of
wealth and wellbeing before running into the more analytical part of the method, the causal flow
mapping. Thus in essence it enabled participants to analyse changes in wealth and wellbeing and
causes of these according to their own understanding of these concepts. Since the exercise was done
systematically across the entire sample, it was possible to produce a synthesis that permits identifying
the overlaps with the characteristics included in the household survey. The synthesis is presented
below. The rows that are coloured in blue in the synthesis tables indicate the overlapping areas,
which (except for “peace of mind”) appear to represent the largest percentages of districts and
participants.
Wealth:
61.
45
82 % of men in 23 districts and 79 % of women in 22 districts identified ownership of property (such
as houses, vehicles, factories or land) as significant signs of wealth. 21 % of men in 6 districts and
25 % of women in 7 districts perceived having many children as indicative of wealth, while 25 % of
men in 7 districts and 4 % of women in 1 district (Assin South) believed that providing one’s
children with quality education was a mark of wealth. For 4 % of men in one district each, indicators
of wealth included possession of farm assets (Nanumba North), inheritance (West Gonja), good
health (Suhum), having one’s family abroad (Abura Asebu), access to loans (Kintampo South), being
educated (Ahafo-Ano South), and having the ability to save in a bank (East Gonja). On the other
hand, 4 % of women in one district each stated that financial strength (Abura Asebu), possession of
farm assets (Nanumba North), being educated (Nanumba North), ability to afford a gari processing
machine (Birim central), good health (West Gonja), having family abroad (Pru), ability to lend to
others (Techiman South), and caring for external family members (Ahafo Ano South) shows that an
individual is wealthy. 29 % of males in 8 districts and 39 % of women in 11 districts mentioned the
possession of large farms as features of wealth. 7 % of men from in 2 districts (Central and East
Gonja) also identified good income, and the ability to invest as signs of wealth, while 11 % of
women in 2 districts (Suhum, Tano North) and 18 % of women in 5 districts (Central Gonja,
Nkwanta South, Wa East, West Gonja and Adansi South) believed that employing workers and
feeding one’s family respectively were indicators of wealth. 7 % of men in 2 districts (Techiman
North and Kpando) and 14 % of women in 3 districts (Central and East Gonja and Wa East) found
A short description of the generic change analysis method is provided in Section 6.3.
29
that being influential in the community and in 4 districts (Central Gonja, Nanumba North, Wa East
and West Gonja) that having more wives signified wealth.
Table 4.1.1: Perceptions of wealth by gender and districts
Participatory wealth characteristics
Having properties (e.g. house, car, larger farm, livestock)
Having large farms
Ability to afford education for the children
Having many children
Higher income from harvests
Ability to invest
Influence in society
Having many wives
Being financially strong and independent
Leaving inheritance for children
Good health and having access to health facilities
Having children or relatives abroad
Having more farm assets
Having access to loan facilities
Being well educated
Ability to contribute to community development
Having savings in banks
Ability to lend money to others
Men
Districts
23
8
7
6
2
2
2
2
2
1
1
1
1
1
1
0
1
0
%
82.1
28.6
25.0
21.1
7.1
7.1
7.1
7.1
7.1
3.6
3.6
3.6
3.6
3.6
3.6
0
3.6
0
Women
Districts
%
22
78.6
11
39.3
1
3.6
7
25.0
0
0
0
0
3
14.3
4
14.3
1
3.6
0
0
1
3.6
1
3.6
1
3.6
0
0
0
0
2
7.1
0
0
1
3.6
Wellbeing:
62.
According to 79 % of men in 22 districts and 68 % of women in 19 districts, peace of mind was
found as an important characteristic of wellbeing. Good health and ability to afford hospital bills
were also considered as fundamental signs of wellbeing by 54 % of men in 15 districts and 32 % of
women in 9 districts. In contrast, access to good water (Tano), ability to educate children (Suhum)
and not being indebted to anybody (Central Gonja) were identified by 4 % of men in one district;
while 4 % of women in one district also perceived ability to solve problems independently (East
Gonja), having lucrative jobs (Pru), and respectful children (Adansi South) as an important
characteristic of wellbeing. 11 % of men in 3 districts (Central Gonja, West Gonja and Adansi South)
and 7 % of women in 2 districts (West Gonja and Adansi South) believed that an individual’s ability
to feed the household was a sign of wellbeing, while 18 % of men in 5 districts (West Gonja, Abura
Asebu, West Akim, Ho Municipal and Manpong) and 14 % of women in 4 districts (Adansi South,
Kintampo South, Techiman North and Agona East) mentioned being able to care for others as a mark
of wellbeing.
Table 4.1.2: Perceptions of wellbeing by gender and districts
Participatory characteristics of wellbeing
Having peace of mind
Good health and having access to health facilities
Ability to care for others
Ability to feed the household
Being financially strong and independent
Ability to solve problems independently
Having respectful children
Having educated children
Having access to good water
Not being indebted to anybody
Having properties (e.g. house, car, larger farm, livestock)
Having lucrative jobs
Men
Districts
%
22
78.6
15
53.6
5
17.9
4
3.6
3
10.7
1
3.6
1
3.6
1
3.6
1
3.6
1
3.6
5
17.9
0
0
Women
Districts
%
19
67.9
9
32.1
4
14.3
2
7.1
2
7.1
1
3.6
1
0
0
0
0
0
0
0
0
0
1
3.6
30
4.1.2 Statistical categories of ‘wealth & wellbeing’
63.
In the household survey, poverty status was determined by assessing each household on its wealth
and wellbeing characteristics for which a proxy means test was used. This test is considered the best
way to assess and determine the poverty status in a statistically relevant manner. Poverty status is
defined broader than ‘wealth’ as purely income and valuable household assets, and also includes
aspects of ‘wellbeing’ such as health and education. The proxy means test uses a scoring formula to
identify categories of wealth and wellbeing by looking at various household characteristics that are
considered important proxies in a particular context. For the Ghana context, these were defined with
assistance of the Ghana Statistical Service (GSS).
64.
The obvious advantage of a proxy means test is that good predictors of wealth and wellbeing such as
socio-economic data, demographic data, housing characteristics, and ownership of household
durables/assets are easier to collect and verify, and provide a more complete picture of the household
status, than direct measures of income. In Ghana for instance, people are generally reluctant to
release their income figures or often don’t know the exact figures. Hence proxies are needed to
measure income as the basis for determining a household’s poverty status, for instance expenditures
per day observed during a certain period of time, which in a brief household survey is not feasible.
65.
Applying a statistical computing procedure known as Principal Components Analysis (PCA) 46,
ownership of household assets and household demographic data collected by the survey47 were used
as proxies and scored to determine the wealth and wellbeing categories in which the surveyed
households were placed. Each household asset for which information was collected was assigned a
standardized score generated through the PCA. The score differed depending on whether or not the
household owned that asset. In the case of education, households were classified first into ‘ever
attended’ and ‘never attended’, and then, for the attended, further into ‘basic’ and ‘secondary’. The
scores were then summed for each household. The resulting asset scores were standardized in
relation to a standard normal distribution with a mean of zero and a standard deviation of one. These
standardized scores were then used to create the break points that defined the quintiles, and the
sample was divided into 5 population quintiles. The first two quintiles (lowest and second) were
reclassified as the category of the ‘poorest’ households, the middle as the ‘less poor’ and the last
two quintiles as the ‘better off’ households.
4.2 Household survey respondent distribution
4.2.1 Gender and poverty distribution in the household survey
66.
The main respondent in the household survey was the head of household. In the random household
sample, 76 % of the household heads were males.
67.
Gender distribution of households among the regions shows that in the Northern region 90 % and in
the Western region 88 % of the household heads were males, while in Central region this figure came
down to only 68 % and in the Brong Ahafo region to 62 %. The proportion of female-headed
46
Principal component analysis (PCA) involves statistical pattern analysis of a dataset, using a technique that emphasizes variation
and helps identify the strongest variation patterns. It does so by identifying values or categories of variables that are linearly
uncorrelated, based on the analysis of data on all possibly correlated variables. These categories are called “principle components”.
More can be found about the method in the specialized literature –e.g.: Abdi, H. & L.J. Williams (2010), Principle Component
Analysis. In: Wiley Interdisciplinary Reviews: Computational Statistics, Vol 2, Issue 4, pages 433–459; Jolliffe, I.T. (2002; 2nd Ed),
Principal Component Analysis. Springer; and open source articles such as
http://pubs.rsc.org/en/content/articlehtml/2014/ay/c3ay41907j
47
See Annex 8 for the household survey questionnaire.
31
households in the Northern Zone was 14 %, which was less than the average of 25 %. According to
the last GSS senses, 35 % of all household heads in Ghana are female; in rural Ghana the proportion
is 31 % (GSS, May 2013).
Table 4.2.1: Zonal distribution of households by gender of the
household head
Sex of Household head
Male
Female
Total
N
Northern
86.4
13.6
100.0
184
Central
71.2
28.8
100.0
424
Southern
74.7
25.3
100.0
229
Total
75.5
24.5
100.0
837
Zones
68.
Of the 837 households surveyed, 335 (or 40 %) were classified as ‘poorest’, 162 (or 19.4 %) as ‘less
poor’ and 340 (or 40.6 %) as ‘better off”. The distribution of the households according to ‘poverty
status’ (as defined by the multi-variant analysis described in Section 4.1.2) is presented Table 4.2.2
below.
Table 4.2.2: Distribution of households by poverty status
Wealth Status
Poorest
Less Poor
Better Off
Total
69.
Frequency
335
162
340
837
Percent
40,0
19,4
40,6
100,0
Of the 335 households that were classified as ‘poorest’, 102 were female-headed (30 %) compared to
233 male-headed (70 %). Of the 162 ‘less poor’ households, 31 were female-headed (19 %) while
131 were male-headed (81 %). Last, of the 340 households that were found relatively ‘better off’, 72
were female-headed (21 %) and 268 male-headed (79 %). Hence female-headed households were
found more in the poorest category, while male-headed households occurred more in the ‘less poor’
household category. Tables 4.2.3a and 4.2.3b show the relative distribution of households per
poverty category and per zone for male-headed and female-headed households respectively. More
detailed poverty distribution tables are attached in Annex 7.
Table 4.2.3a. Poverty distribution of the Male households by Zone
Zone
Northern
Central
Southern
Total
Poorest
45
126
62
233
Poverty Status
Less Poor
33
64
34
131
Table 4.2.3b. Poverty distribution of the Female households by Zone
Poverty Status
Zone
Poorest
Less Poor
Northern
2
7
Central
71
12
Southern
29
12
Total
102
31
Better Off
81
112
75
268
Better Off
16
39
17
72
Total
159
302
171
632
Total
25
122
58
205
32
4.2.2 Distribution of livelihood sources
70.
From the household survey, it is clear that R&T production is the first most important source of
income with little variation across poverty status (61 % poor, 83 % non-poor and 77 % richer). R&T
processing is less popular (25% poor, 4 % non-poor and 7 % richer households) yet adds to the total
percentage of households that depends on R&T as a primary source of income. Across all the
identified socio-economic categories, more than 80 % of the households have R&T as their most
important livelihood source. This does not come as a surprise since the households were sampled to
include 60-70% intended program beneficiaries to enable co-variant analysis and linking to the
findings produced by the other PIALA methods with regards to R&T production and processing,
while 30-40% were sampled from other community members to inquire the program’s assumption
that improvements in R&T-based livelihoods would also affect all other people living in the
catchment areas. In many locations, however, most (if not all) households turned out to be intended
beneficiaries, having R&T as an important source of income.
Figure 4.2.1: % distribution of the first most important source of income for households.
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
Production of
R&T
Processing of
R$T
Transportation
of R&T
Poorest
61.2
24.5
0.0
Other
livelihood not
R&T
14.3
Less Poor
83.3
3.7
0.6
12.3
Better off
76.8
7.4
0.6
15.3
71.
Unlike the first most important source of income, the second one does vary with poverty status.
For 28 % of the households that are presently considered poor, the second most important source is
other livelihood activities (thus not R&T-related), while 31 % of these households does not have a
second livelihood activity. For 40% of the non-poor households and 38 % of the richer households,
R&T-production remains central as a second most important source of income, while 33 % non-poor
households and 31 % richer households have other livelihood activities as their second source. Also
15 % of non-poor households and 19 % of richer households have R&T processing as a second most
important source of income.
72.
In 44-46 % of the households, the two most important sources of household income are controlled by
the husband, compared to 22-29 % of the households where the wife is in control of these sources.
73.
Households that depend on merely one source of income are more at risk of hunger and poverty if
that source comes under pressure. Only 17 % of the surveyed households depend on a single source.
The majority 49 % has three or more sources. This seems to confirm the positive trend of increased
access to sufficient food mentioned earlier.
Figure 4.2.2: % distribution of the total number of sources of income for households
Three main
sources of income
49%
only one main
souce of income
17%
Two main sources
of income
34%
33
4.3 Distribution of participatory research participants
74.
The participatory research of the supply chains involved a total of 109 gender-specific FGDs
(53 with women and 56 with men), in which 839 community members participated (411 women and
428 men; and 90 % intended beneficiaries) and 43 mixed FGDs with in total 341 participants (179
women and 162 men). In addition, 75 KIIs were conducted with district officials and service
providers (including private actors involved in the program, such as the leaders of GPCs, other SMEs
and the local branches of the PFIs). Of these FGD and KII participants, average 28 were invited to
participate in a half-day sensemaking workshop organized in every district at the end of the data
collection. In total 23 workshops were held, engaging 640 research participants, of which 81 %
intended beneficiaries (48 % female and 52 % male farmers and processors). Table 4.3.1 below
provides an overview of the supply chain research participants. The more detailed distribution tables
of the FGDs according to methods can be found in Annex 4.
Table 4.3.1: Distribution of Supply Chain Research Participants
Type of method
KIIs with district officials
KIIs with service providers
Generic Change Analysis with RTIMP intended beneficiaries
Livelihood Analysis with RTIMP intended beneficiaries
Constituent Feedback with DSF participants
Constituent Feedback with FFF participants
Constituent Feedback with GPC participants
District Sense-Making Workshops with RTIMP intended
beneficiaries, participants, district officials and service
providers
F
%
M
%
222
189
42
58
79
51
47
49
41
69
217
211
43
84
35
49
53
51
59
31
N
36
39
439
400
85
142
114
640
4.4 National respondents and participants
75.
Following a desk review and consultations with the program coordination unit, a design workshop
was organised on 12 October 2014 in Kumasi to engage the evaluation commissioners and national
stakeholders in important design decisions for framing and focussing the evaluation (including
finalizing the ToC, determining the assumptions and evaluation questions to be inquired, deciding on
the sampling populations, and agreeing on the evaluation rating system). The participants in this
workshop were invited to further take part in the evaluation’s Core Learning Partnership (CLP).
These included: the national and zonal RTIMP coordinators and senior staff, members of the RTIMP
Steering Committee, the IFAD country program manager and senior staff, IFAD’s supervision and
evaluation consultants, MoFA national and regional directors and officers, and representatives of the
participating financial and research institutions, the SCFs and TREND. The major outcome of the
evaluation design workshop was the design paper48.
76.
As part of the evaluation design and data collection, about 25 interviews and consultations were held
with RTIMP and IFAD officials, participating financial institutions, FFF research leaders, off-takers
and other service providers. An overview of all the interviews conducted with stakeholders at local,
zonal and national levels is provided in annex stakeholders interviewed provided in Annex 3.
48
Cf. Van Hemelrijck, A. & G. Kyei-Mensah, 2014.
34
5
Field research methodology
77.
Nested mixed-methods were used for collecting quantitative and qualitative evidence concurrently
and independently, with equal weight, to investigate the causal links in the ToC. Quantitative survey
and individual scoring with randomly selected primary and secondary program beneficiaries were
combined with qualitative causal flow mapping using participatory processes and recall and
triangulation techniques to inquire the evaluation questions.
78.
All the methods, tools and questionnaires used in this impact evaluation have been field-tested and
adjusted to the Ghanaian context. A detailed field manual was put together for the researchers to
ensure that the methods would be used and data collected in an appropriate and systematic manner
across all populations, and that participatory processes would be facilitated in a way that is sensitive
to power dynamics, inclusive, ethical and free from external influence. Standard note-taking
templates and data entry spread sheets were developed and used to warrant systematic data capturing
and early data processing. Raw data reports were produced on all FGDs, interviews, workshops and
desk review.
5.1 Key Informant Interviews
79.
Over 100 semi-structured interviews were conducted with national, zonal and district-level program
stakeholders. At the national and zonal level these included RTIMP and IFAD officials, managers
from the PFIs, the FFF research leaders, and a few important off-takers or industry leaders. At the
district-level these included district officials, leaders of GPCs and other SMEs, and the managers of
the local branches of the PFIs. The KII questionnaires are attached in Annex 9.
5.2 Household survey
80.
To measure the impact of RTIMP in terms of changes in access to food and income, a succinct
household survey was conducted with 837 households in the 25 sampled districts. Through this
household survey, essential data was collected on changes in food, assets, income, R&T revenues
and R&T activity. The household survey questionnaire is attached in Annex 8. The results from the
correlation analysis can be found in Annex 1.
81.
The data was captured with the Census and Survey Processing System (CSPro), which is a software
package for entering, editing, tabulating, and disseminating data from censuses and surveys. CSPro
combines the features of the Integrated Microcomputer Processing System (IMPS) and the Integrated
System for Survey Analysis (ISSA). CSpro is a data entering application developed by the U.S.
census bureau, ICF international and Serpro S.A. Data cleaning was also done using the CSpro.
Clean data generated by the CSPro application was then exported into a SPSS database system for
analysis. While CSPro provides some tabulation capabilities, it is not intended to replace more
sophisticated statistical analysis software such as SPSS, Stata, etc. All the analysis was done in
SPSS. The tables generated in SPSS were exported to Microsoft Excel 2010 for the preparation of the
charts and graphs.
5.3 Generic change analysis
82.
Complementary to the household survey, gender-specific focus group discussions using the generic
change analysis method collected qualitative data on improvements in livelihoods that affected
wealth & wellbeing, in order to capture not only intended but also unintended influences (both
35
positive and negative) on rural poverty. For this it used 2
PRA-based tools, namely: change ranking and causal flow
mapping. The change ranking is a descriptive data
collection tool that sought to identify and rank the main
changes in roots- & tubers-based livelihoods of the past 5
years in terms of their impact on people’s wealth &
wellbeing as defined by the beneficiaries themselves49.
Subsequently the causal flow mapping inquired the possible
explanations by taking the one or two changes with greatest
impact (thus highest rank) as a starting point to map out
their impacts and causes, link these back to RTIMP, and
collect detailed information on who had been affected
most/least and why. The guidance note for the generic
change analysis method can be found in Annex 10.
5.4 Livelihood analysis and SenseMaker lithe
83.
To further investigate R&T livelihood changes and causes that affected household food and income,
focus group discussions using the livelihood analysis method were held separately with men and
women (as in the generic change analysis). This method combines two PRA-inspired tools (change
matrix and causal flow mapping) and a small SenseMaker50 experiment. The change matrix is a
descriptive data collection tool that helped to obtain an overview of the different types of livelihood
activities in the communities related to roots and tubers and the major changes that happened in these
livelihood activities in the past 5 years, as well as women’s and men’s engagement in each of these
and the relative income and risk levels. For this it used PRA-based techniques such as ranking,
proportional piling and scoring. The causal flow mapping was an explanatory data collection tool
similar to the one used in the generic change
analysis method, that helped mapping out the
impacts and causes of the one or two most
significant changes in the R&T livelihood
activities, link these back to RTIMP, and collect
detailed information on who had benefited (or
not) and why.
84.
SenseMaker was employed in this evaluation in
a lithe version. Individual experiences (393 of
which 246 positive and 147 negative) were
collected from the participants in the livelihood analysis FGDs, which were no longer than one or
two lines (instead of full-fledged story which is usually the case in a normal application). The
experiences were self-signified by the participants using 5 basic tools (2 of which related to the
impact cluster, 2 to the processing and the production clusters, and 1 to the market-linking cluster in
the RTIMP ToC). Patterns in these experiences were then analyzed using the SenseMaker software.
The purpose of this lithe SenseMaker experiment was to test and demonstrate its potential value and
identify the methodological challenges when used for impact assessment. Complementary to the
household survey and the PRA-based tools, it helped surface patterns of perceptual change, impact,
49
50
A synthesis of the characteristics of wealth & wellbeing as defined by the research participants is presented in Section 4.1.1.
SenseMaker is a patented approach of Cognitive Edge (cf. http://www.sensemaker-suite.com).
36
causes and influences, and provided an
additional layer of quantified qualitative data
collected from a relatively large population.
85.
The guidance note for the livelihood analysis
method, including the SenseMaker tools, is
attached in Annex 11. There is also a separate
sub-report on the findings from the
SenseMaker analysis.
5.5 Constituent feedback
86.
Last, the constituent feedback was used to collect quantified perceptual data on the reach and effects
of the evaluated program mechanisms (DSF, FFF and GPC/MEF) on R&T livelihood changes and
causes, from the perspective of the beneficiaries or constituents. Constituent feedback (also called
constituent voice 51) is a low cost performance monitoring method for collecting quantified
qualitative feedback and engaging in dialogue with key constituents or beneficiaries, using
standardized metrics similar to the customer satisfaction surveys developed in the private sector. It is
mostly empowering and effective for improving performance when used recurrently throughout the
lifetime of a program. The guidance note that was used for the constituent feedback FGDs and
scoring can be found in Annex 12.
87.
For each of the program mechanisms, a focus group discussion was organised with the intended
direct beneficiaries around a small number of questions, which involved an individual and
anonymous scoring by each of the participant. The scoring was done behind the back while the
facilitator goes around and takes notes of the scores Some of the questions were also asked in the
KIIs with the service-providers to mirror their
views with those of their clients52.
88.
Similar to the SenseMaker experiment, the
purpose of using constituent feedback in this
impact evaluation was to test and demonstrate
its added value and identify the methodological
challenges when used for impact assessment.
The findings from the pilot testing of these two
methods are presented in a separate report on the
PIALA methodological reflections.
5.6 Data consistency and quality monitoring
5.6.1 Data capturing and collation
89.
Qualitative and quantitative data were collected efficiently, systematically and ethically by using the
same set of methods and tools for facilitation, interviewing, reflection, sensemaking, data capturing
and data storing in a systematic way in all localities.
51
Cf. http://www.keystoneaccountability.org/analysis/constituency.
Including: the FFF facilitators;the 7 research team leaders from CSIR, KNUST & UCC; the DDAs, BACs, SCFs and RTIMP
desk officers; the directors/leaders of the GPCs and other supply chain leaders, and of the PFIs.
52
37
90.
To avoid data loss or contamination, given the huge quantum of data that was generated in each
district, clear note-taking instructions were followed and data were entered and sent to the head
office of the research firm in Accra every day.53 Note taking was done as accurate as possible,
avoiding changes in the wording or researchers’ interpretations of any of the responses. To ensure
consistency needed for the aggregated analysis, standardized data capturing templates and spread
sheets were used. To enable early (almost instant) data processing and crosschecking during
fieldwork, a standardized data collation tool and format were employed that helped the researchers to
integrate and connect all data to the causal links in the ToC. The data collation tool can be found in
the guidance note on data collation & quality monitoring attached in Annex 13.
91.
A simple rating system was used for estimating the robustness of the emerging evidence, the extent
and scope of each causal link, the extent and quality of program contributions, and the interference of
other influences. Filling in the table systematically using this rating system enabled the teams to
timely identify data gaps and weaknesses, and prepare for the district sensemaking workshops. Also
the data collation tables and district sensemaking workshop reports were sent to the head office in
Accra at the end of the fieldwork in every district. All this was closely supervised while technical
backstopping was provided at any moment in time, wherever there was doubt or the teams asked for
assistance.
5.6.2 Participatory research ethics and independence
92.
It is very important to ensure that research is both ethical and accurate. To warrant this, ethics was a
key part of the field research training for the entire team involved in this piece of work. All
researchers conducted their activities as independent, ethical and accurate as posible. They were very
mindful of their task to collect reliable data and made it a point to maintain independence from
RTIMP influence in the sampling and the data collection.
93.
Although district and RTIMP officials played an important role by giving details of the project and
introducing the team to the communities, most of the commuity mobilization was carried out by the
researchers, and all FGDs and household surveys took place without any interference or presence of
officials. Also most officials and community leaders were not aware of the evaluation until the
moment that the teams entered the district. The teams moved fast and left barely any time for
officials to notify and influence the research participants.
94.
Focus group participants were selected
without exclusion on the basis of, for example,
class or stigma while cultural and community
norms were understood and considered in the
selection process. Focus groups were
disaggregated by gender to avoid the
traditional power dynamics between men and
women that are common in traditional WestAfrican communities. Two team members
were present in each FGD – one facilitated
while the other took notes. At the start of
every FGD, purpose and limitation of the research were clearly communicated to the participants.
The teams also ensured that demands on participants’ time were not excessive and they were aware
of their right to not participate or withdraw at any time.
53
When there was no internet available, data was sent by mobile phone.
38
95.
Data collection (HH surveys, KII and FGD) was facilitated in such a way that respondents were
made to feel at ease and encouraged to equally ask the researchers questions off their own. The team
recognised that participants were possibly vulnerable, hence all exercises were facilitated in a way
that was highly sensetive to participants needs and fears and as much as possible tried to ease power
differentials that occurred among community members and between participants and researchers.
Research questions were asked using simple local language. Team members all spoke several local
languages. In the few cases where translators were needed, they were of the same gender and status
as the respondents.
96.
All interviews and discussions took place in private spaces free from people with power or authority,
or people who could intimidate or influence the responses. Interviews and FGDs with resource-poor
farmers and processors did not have local leaders or officials from RTIMP or the GPC present at any
moment in the process. Anonymity and confidentiality were important hallmarks during the entire
field research –also during record keeping and report writing. Researchers made sure that
participants understood that what they did and said in the group sessions would remain entirely
anonymous.
97.
Meetings that invoved local leaders and officials of RTIMP were carefully facilitated with due
attention paid to power dynamics. For example, in the district sensemaking workshops, farmers and
processors were always given the chance to discuss things with their peers first and then present their
critical comments before any of the officials or service providers were given the opportunity to speak.
5.6.3 Reflective practice and quality monitoring
98.
Data quality monitoring involved daily research team reflections on research processes and outcomes
as the basis for timely identifying data gaps/weaknesses and assessing the robustness of the evidence
base being built. Data quality monitoring went hand in hand with data collation, and involved
average a 2-hour process every evening to discuss the team’s observations on the process by which
the data had been collected during the day, and the conditions and power issues that could have
influenced the data and potentially generated a biases.
99.
Five key sets of questions were provided to the teams to guide the evening reflections and help them
assess the robustness of their evidence. These can be found in the guidance note for data collation
and quality monitoring attached in Annex 13. The leading questions for these 5 sets were:





Were we able to conduct the interviews and FGDs as planned and in a systematic manner?
How well did we facilitate discussions to ensure active engagement and equal voice?
Were we able to properly code and document all data?
Were we able to obtain robust and sufficient data on the causal links in the ToC?
Were we able to collect sufficient data on the value and reach of specific program
mechanisms?
5.7 Participatory sensemaking
100. Normally not included in impact evaluation, yet an important part of the PIALA, was the 2-stage
participatory sensemaking process that was organised to engage all stakeholders (including
beneficiaries) at the local and national levels in a collective analysis and discussion of the evidence in
relation to the links in the ToC. Going beyond simply validating evidence by also involving
stakeholders in a collective analysis is quite new in the field of impact evaluation. Traditionally, this
is considered part of the dissemination and learning that comes after the evaluation. Obviously, if
done after the evaluation, it doesn’t leave room for feedback or build ownership of the evaluation. As
39
a result, evaluations often remain unused. Doing this before finalising fieldwork in every locality
and before finalising the aggregated analysis at the national level helps improve and strengthen the
evidence base, while avoiding top-down data extraction and researcher-dominated analysis.
101. The purpose of participatory sensemaking is to enhance the empowering value of impact evaluation.
It does so by creating a space where program stakeholders can freely and collectively discuss the
evidence, and by providing the tools that help them look at development processes from a systemic
perspective and value program contributions to impact among other influences as part of a bigger
picture. For this to succeed it is essential that all participants are enabled to critically engage with the
evidence, have an equal voice and feel comfortable to express their views. This is quite challenging
when bringing policy makers, service providers and intended beneficiaries (who are mostly illiterate)
in the same room. A participatory sensemaking workshop model has been developed for this that
builds on 8 design principles and uses concepts and tools of reverse engineering, active listening,
patches & nodes, iterative process, different vantage points, soft systems modelling, and contribution
analysis. First piloted in Vietnam, this model was further expanded and improved in this impact
evaluation.
102. Using this model, sensemaking workshops were
organised at two levels: locally and nationally.
At the local level, the process was organised
around the main unit of analysis: the supply
chains that were administered at the district
level. In total 23 half-day district sensemaking
workshops were organised with an average of
28 participants per workshop, of which 81 %
intended beneficiaries (48 % female and 52 %
male farmers and processors). The total number
of participants in the district-level sensemaking
process was 640. At the national level, the process was centred on the aggregated analysis of
program contributions countrywide. A two-day sensemaking workshop was organised on 6-7 May in
Kumasi, involving 106 participants, of which 40 % intended beneficiaries (38 % female and 62 %
male farmers and processors, many illiterate), 45 % local and national officials, and 15 % private
sector actors. The process was quite successful in creating a lively debate and helping participants
understand the program and its influence on livelihoods and poverty systemically. The guidance
notes for the local and national sensemaking workshops are attached in Annexes 14 and 15. The
textbox below presents the debrief that was sent to IFAD in Rome right after the workshop.
Textbox 6.7.1. Debrief on the PIALA national sensemaking workshop (8 May 2015)
We opened the workshop with the phrase: “The
reason why you are all invited here is not to just
talk about RTIMP, but to discuss how we can build
on the RTIMP and do greater things in the next big
program, the GASIP”, followed by a quick exercise
in which we let people first think of what to do if
they were in the position of government and needed
to better reach out, make policy, plan or get
feedback in order to improve rural lives and
livelihoods, and how to use learnings from impact
assessment. This opening assignment formed the
reference point to which we returned several times
during the workshop.
40
The rest of the day we ran into focus group discussions on evidence statements that the researchers had
prepared based on the analysis of the quantitative and qualitative data they had collected in 25 random
districts across the country, and reconstructed the
theory of change with these evidence statements for
the north, the center and the south of Ghana.
People worked in focus groups composed of their
peers, as to make them feel comfortable and safe to
form their own critical judgments. The PDA
research team did an outstanding job at facilitating
these groups.
Only at the opening of the second day we had a
first real plenary discussion in the form of a "fish
bowl" in which the participants discussed their
ideas about areas that have not been successful in
the RTIMP and require more effort in the
GASIP. The fish bowl consisted of 9 chairs in a
small circle (3 for the farmers, 3 for the processors, and 3 for the officials, bankers and enterprises). Only 2
of each group was allowed
to take the seats, so each time a new participant took one of the empty seats, another of his/her group had to
leave. The bulk of the discussion was about issues of developing businesses, accessing credit and linking to
markets. This worked surprisingly well, despite the fact that most of the participants were farmers and
processors who could not read or write.
The rest of the morning of the second day was used
for rating RTIMP contribution in focus groups organized per zone and at national level. The rating was
done for the evidence statements selected by each group as most important, on a canvas with X-axis
“positive - negative” and Y-axis “strong RTIMP influence - RTIMP total overrun by other influences”.
Interesting was to see how different ratings came out for the same statements. Most astonishing was the
rating of ALL statements as totally positive and
entirely attributable to RTIMP by the national
group, as opposed to the groups of processors and
farmers who rated mostly somewhere in between
and some even negative and with little RTIMP
contribution. It would be interesting to do this
during a program’s lifetime as part of for instance
an annual impact reflection using evidence from a
regular RIMS outcome survey that uses some of the
PIALA methods.
After the farmers and processors had left, we
continued in the afternoon discussing the design of
the evaluation, the methodology, the key findings,
and the potential for integrating the approach in the
next GASIP with the national stakeholders. The first part of this discussion was about the robustness of the
approach and the cost-benefits (what PIALA can or cannot do, and what it could do more if used as an
integrated approach linked to program M&E). The last part was about the findings, and particularly new
insights useful for the GASIP.
5.8 Methodological strengths and limitations
103. The first PIALA pilot in Vietnam experienced several limitations from which much was learned in
the adjusted approach employed in the second PIALA pilot in this impact evaluation of RTIMP in
Ghana. Issues of sampling related to the heterogeneity in program distribution and treatment,
political influence and organisational pressure in the participatory processes, and systematic data
collation and quality monitoring during fieldwork to ensure data integration, were adequately
addressed in the evaluation in Ghana. This resulted in substantial improvements in the quality of
evidence.
41
104. Research capacity for appropriately employing PIALA was indispensible in this regard. PDA (the
research firm that was commissioned to conduct the impact evaluation) was well equipped to manage
a complex and countrywide methodological innovation and evaluation project. It had two offices:
one in the capital (Accra) and one in the centre of Ghana (Kumasi), both well staffed with
administrative and logistical assistants (incl. drivers) and researchers competent in organising
fieldwork and workshops of different types. Researchers and logistical staff had extensive knowledge
of road networks and geography in Ghana, local cultures and local languages54, and the requisite
skills for community mobilization. Researchers were experienced facilitators of participatory
processes, well versed in handling group dynamics sensitive to the needs of local communities,
officials and development professionals, and capable of rigorously processing large amounts of
qualitative and quantitative data. PDA also had the ability to contract independent researchers and
consultants for those parts of the research that required specialised or higher level expertise. A senior
statistician from the Ghana Statistical Service (GSS) was contracted to lead on the sampling and the
household survey.
105. The major strengths of this RTIMP evaluation were: (a) the selection and use of methods specific to
the causal links in the ToC and the evaluation questions; (b) the comparative analysis of the relative
contribution to impact of heterogeneous configurations of program treatment (as an alternative for a
classic counterfactual analysis); and (c) the 2-stage participatory sensemaking process that engaged
all stakeholders, including beneficiaries, in a collective analysis and discussion of the evidence.
Alongside these strengths, there were also some challenges and constraints encountered by the
research team in the conduct and management of this evaluation. Three key constraints requiring
more attention in future evaluations using PIALA are: (a) the sampling of market-bounding systems
such as supply chains centred around supply chain leaders, which per definition have open
boundaries and thus are difficult to discern, particularly when interacting and thus overlapping in the
same geographic and administrative location; (b) the time and capacities required from the people to
participate in FGDs using PIALA methods, in particular when most are illiterate (e.g. the use of pen
and paper or even tablets, the length of the FGDs, SenseMaker tools using abstract concepts, etc.);
and (c) the rigid nature of the methodology that needed to be applied in a systematic manner across
all locations, which sometimes clashed with the cultural settings in some communities and was
difficult to maintain in the limited time that was spent in each district. Also the limited access to
power in remote areas and the frequent power interruptions formed a major problem to keep up with
the tight field research schedule. The researchers often had to stay at rural guesthouses that did not
have generators, which implied there was no light to type out field notes to be sent to the research
manager and do the evening reflections and data collation as planned.
106. The main take-away for future PIALA applications is that (a) methods and tools need to be adapted
to the participants’ conditions as much as possible, and (b) sufficient time is needed in the field to
accommodate cultural habits and events and address unexpected challenges with regard to sampling
and mobilisation. Obviously, if PIALA methods and tools would be used regularly as part of ongoing
M&E, then this would certainly help overcome these differences and challenges and contribute to
building participants’ capacities and empowerment. Moreover there would also be a much bigger
return from the resources (time and budget) invested in the training of an independent research team
in PIALA and the focussing and framing of the program’s impact assessment and learning agenda
using PIALA as an overarching approach. This is discussed in greater detail in a separate report on
the PIALA methodological reflections.
54
Ghana has over 80 different languages. Most researchers spoke 3-5 languages, a few spoke more than 10. An advantage was that
also the logistical assistants and drivers spoke a good number of Ghanaian languages and dialects.
42
6
Configuration analysis
6.1 Analysis of causal links
107. The configuration analysis of the evidence collected in the 25 districts consisted of three important
steps. First, all evidence from the district data collation tables were entered in an aggregated data
collation table55. Second, all scores and explanations entered in that table were cross-checked on
consistency and adjusted where needed based on a thorough reading of the original field reports and
deliberations with the lead research team. Third, a comparative analysis was conducted of the
similarities and differences in the evidence for all the causal links in the ToC, and the effects on
livelihoods of the various patterns or configurations of program treatment emerging from the analysis.
For this a combined system was used of: (a) binary coding of the formal presence of program
mechanisms, and (b) scoring of the quality, strength and consistency of each of the causal links in
which the mechanism operate.
108. The analysis was done by tracing the cascading changes and causes in the ToC back from the impactlevel to the level of the program mechanisms, and looking for major similarities and differences
between districts across the three agro-ecological and administrative zones. Starting with the impact
claim, districts were first clustered according to their scores on improvements in R&T livelihoods.
Differences and similarities were examined within and between these clusters. Next, findings from
the analysis of impact-related changes and causes were then used to examine the changes and causes
in each of the contributions claims separately, using a similar procedure as the one for the impact
claim. Districts were clustered according to their scores on the enhancement of market linking for the
first claim, enhancement of production for the second claim, and enhancement of processing for the
third claim. Differences and similarities were then examined within and between the different clusters
for each of the three claims.
109. The findings from the analysis for the impact claim are presented in Section 7, while the findings
from the analyses for the three contribution claims can be found in the subsequent Sections 8-10.
6.2 Scoring of causal links
110. A simple rating system was used for scoring the causal links in the ToC on their relative strength and
consistency based on strength and quality of evidence. The rating system is similar to the one used by
IFAD in supervisions and reviews, but adjusted to serve the purpose of assessing RTIMP’s
contributions to impact. In this system, score 6 rates ‘highly satisfactory’ meaning that the
contribution is very robust and attributable to RTIMP, score 5 ‘satisfactory’ thus strong and
attributable, score 4 ‘moderately satisfactory’ which is strong but inconsistent or not entirely
attributable, score 3 ‘moderately unsatisfactory’ or weak but consistent thus reasonably attributable,
score 2 ‘unsatisfactory’ or weak and inconsistent so barely influenced by RTIMP, score 1 ‘highly
unsatisfactory’ which means there is no or even a negative RTIMP influence, and finally 0
‘insufficient evidence’ implying the evidence is too weak to score contribution. The same rating
system was also used to score the quality of evidence for each link on its relative strength and
consistency.
111. Table 6.1.1 below presents a synthesis of all the scores from the aggregated data collation table for
the three contribution claims and their effects on livelihood improvements. The districts are clustered
according to the score for the link [O1+O2+O3I2] in the ToC, which represents the effects of the
55
All collation tables are part of the raw research materials that are submitted to the evaluation sponsors.
43
RTIMP’s contributions on livelihoods. The supply chain areas or districts are listed in the left column.
The next four columns contain the binary codes indicating the availability of the program mechanism,
in which “0” stands for “not present” and “1” for “formally present”, while “NA” means “not
applicable” in the case of yam that is exported fresh and therefore doesn’t require any processing
centre or GPC. The code doesn’t say anything about the quality, reach and effects of the mechanisms
and thus their influence on the causal links. The formal availability of a certain mechanism doesn’t
imply it was actually accessible. The next nine columns present the scores56 for the main links in
each of the three contribution claims, the accessibility and performance of the mechanisms in these
links, and the quality of evidence. The cells that are shaded in blue reflect the formal absence of the
mechanism. Where cells are shaded blue yet contain the code 1 indicating the formal presence of a
GPC, this means the GPC was indeed formally there, but not operational.
56
A justification for the scores can be found in the aggregated collation table.
44
Table 6.2.1. Rating of RTIMP contribution to the improvement of R&T-based livelihoods
Gari
NZ =
CZ =
SZ =
Yam
PCF
HQCF
Other
Northern Zone
Central Zone
Southern Zone
DSF FFF GPC MEF
Tano North (Apesika) (CZ)
Techiman
(CZ)
Gomoa East
(SZ)
Assin South
(SZ)
Birim Central
(CZ)
Nkwanta South
(NZ)
Upper West Akim
(CZ)
Ashanti Mampong
(CZ)
West Gonja (Damongo) (NZ)
Abura Asebu Kwamankese (SZ)
Nanumba North
(NZ)
East Gonja
(NZ)
Central Gonja
(NZ)
Suhum
(CZ)
Adansi South
(CZ)
Ahafo Ano South
(CZ)
Kintampo South
(CZ)
Wa East
(NZ)
North Dayi/ Kpando
(SZ)
Agona East
(SZ)
Pru
(CZ)
Ho Municipal
(SZ)
Tano North (Dua Yaw (CZ)
Nkwanta)
Wassa Amenfi West
(SZ)
Kumasi Metropolitan (CZ)
Assembly
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
1
1
0
0
0
0
1
1
1
1
1
1
1
1
1
1
(M3c)+C1a+M3b
C3c
1
1
0
1
1
0
1
1
0
1
3
4
2
3
3
3
2
3
3
3
0
1
0
2
3
2
2
0
0
0
2
2
2
0
1
N/A
N/A
N/A
1
1
1
Contribution Claim of
RTIMP Component 3

Enhanced Processing (O3)
MEF
GPC
Evidence
5
5
3
4
4
5
4
5
5
5
3
2
1
1
2
1
N/A
0
1
N/A
N/A
Strength
6
5
5
4
3
4
4
4
4
3
N/A
N/A
3
4
4
2
N/A
2
2
3
N/A
2
2
N/A
0
1
1
(M3b)+C3c C3b
 O3
Contributions of
RTIMP Component 2
RTIMP Component 1
RTIMP Components 1, 2 & 3



Enhanced Production (O2) Enhanced Market-Linking (O3) Improved Livelihoods (I2)
FFF
DSF
Evidence M1c+M1b+
Evidence O1+O2+O3
Evidence
O3+O2
M2a+M2b+ C2a+C2b O2
Strength
C1a+(M1) O1 Strength
Strength
+O1C1b
I2
(M2c) C2a
5
4
5
5
5
6
5
5
5
5
5
5
5
4
5
4
4
5
4
4
2
4
2
5
3
5
5
5
5
5
4
5
5
4
5
5
3
4
4
3
4
4
5
3
4
2
5
3
5
5
5
4
4
5
4
5
5
6
5
5
5
5
4
5
3
5
5
4
4
5
4
4
4
4
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
2
2
3
2
4
4
4
3
4
3
3
3
3
3
3
3
2
3
2
2
3
3
2
2
2
3
2
5
5
5
4
4
5
5
5
5
5
5
5
5
5
3
5
5
5
5
6
5
5
3
5
5
5
4
4
4
4
4
4
4
4
4
4
3
3
3
3
3
2
2
2
2
2
5
5
6
4
5
5
5
5
5
4
5
5
5
5
5
5
5
5
5
6
5
5
3
5
4
4
1
4
1
5
4
1
1
1
1
6
4
2
1
5
4
5
5
5
4
5
6
6
45
7
Findings on impact-related changes and causes
112. This section presents the key findings from the analysis of the evidence collected on the causal links
(changes and their causes) in the impact claim of the ToC of RTIMP. For this evaluation, impact
was defined in terms of changes in “access to food and income for the rural poor to sustain an active
and healthy life” resulting from improvements in R&T-based livelihoods in the catchment areas of
the four main commodity chains (namely: Gari, HQCF, PCF and FYE) that have been developed
with support from RTIMP in the past five years (2010-2015).
113. The assumption was that livelihoods and poverty status could be improved by commercializing
smallholder R&T production and processing businesses combined with the development of
competitive market-driven and inclusive supply chains. The main question that the evaluation sought
to answer was to what extent this assumption held true (or not) under which conditions. To answer
this question, two causal links were assessed as part of the impact claim:

I2->I1 which is about changes in access to food and income among
the rural population as a result of changes in R&T livelihoods;

O1+O2+O3->I1 which is about changes in R&T livelihoods as a
result of changes in market-linking, production and processing.
114. The findings related to the 2 causal links in the impact claim are described
and explained in greater detail in the next two Sections 7.1 and 7.2.
O2
I1
I2
O1
O3
7.1 Changes in access to food & income and its causes
(link I2I1)
7.1.1 Changes
115. Although the 2008 baseline and other studies such as the 2009 CFSVA showed an already high
average of food sufficiency at the time, five years down the road of RTIMP, the household survey
indicates an undeniable and significant increase in access to sufficient food in the RTIMP treatment
areas, which cuts through all socio-economic household categories across the entire country.
116. RTIMP baseline figures showed that average 85 % of the in 2008 surveyed households living from
the production, trading and processing of roots and tubers could feed themselves throughout the year.
The WFP’s Comprehensive Food security & Vulnerability Analysis (CFSVA) found in 2009 that
95 % of Ghana’s entire population had access to sufficient and nutritious food for leading an active
and healthy life and thus being defined as ‘food secure’.
117. Statistical analysis of the household surveys showed that more households have gained access to
sufficient food throughout the period of 2014 than in 2013 across all socio-economic household
categories. Figure 7.1.1 below presents the distribution of households who experienced food shortage
classified by relative poverty status of households. For instance, 80 % of poor households did not
experience any food shortage in 2014 compared to 57 % s in 2013. There was not much difference
between female- and male-headed households in terms of the distribution of those who experienced
46
food shortage: 16 % male- compared to 18 % female-headed households in 2014, and 49 % maleversus 46 % female-headed households in 2015.57
Fig 7.1.1: % distribution of households that did not experience any food shortage in 20132014, by poverty status
100.0
80.0
56.7
80.0
89.1
78.4
50.0
46.3
60.0
40.0
20.0
0.0
Poorest
Less Poor
Better off
2014
2013
118. Less positive are the distribution figures regarding households who have put their children under 5 to
sleep at night without a meal, which in Ghana is exceptional and an indication of food shortage or
shortage in cash. Of the male-headed households, 74 % have put their children under 5 to sleep
without a meal at least once in 2014, compared to 71 % of the female-headed households. Compared
to 32 % of the female-headed households, 45 % of the male-headed households have put their
children under 5 to sleep without a meal at least twice in 2014.
119. There has been a noticeable increase in household income levels proportional to an increase in the
annual total value of household R&T production and processing in the RTIMP treatment areas in the
period between 2009 and 2014, suggesting a positive impact particularly in a context where there is
a limited rural infrastructure and an enduring land tenure issue. Increases in household income and
R&T value seemed to have reached a ceiling indicating stagnation in the improvements of R&Tbased livelihoods.
120. Figure 7.1.2 below shows that the proportion of households with annual income larger than 5,000
GHS increased from 10 % in 2009 to 14 % in 2014, and the proportion of households with annual
income between 2,001 and 5,000 GHS increased from 12 % in 2009 to 26 % in 2014. At the same
time, the proportion of households with annual income less than 500 GHS has decreased from 18 %
in 2009 to 12 % in 2014, and with annual income 501-1,000 GHS from 25 % in 2009 to 15 % in
2014. As more households gained higher incomes, more households also gained access to food (p
0.32).
57
The little variance across the poverty categories, and particularly the slightly higher occurrence of food shortage among the less
poor in 2014 and among the better off in 2013, may appear counterintuitive. Yet these categories were determined based on the data
collected from the households on characteristics identified as best proxy’s for defining relative poverty status by means of a
statistical computing method called Principal Components Analysis (PCA). Hence most likely there wasn’t much difference in the
characteristics among the majority of rural households, which is also confirmed by other studies about rural poverty in Ghana. A
more fine-grained set of characteristics requiring a much longer household survey would have been required to obtain more
accurate quintiles and thus categories. The survey though was kept succinct on purpose as to avoid respondent and researcher
fatigue and allow more space for mixed-methods research.
47
Fig 7.1.2: % distribution of the range of total income of the household from 2009 to 2014
30
25
20
15
10
5
0
0-500
501-1000
1001-2000
2001-5000
> 5001
2009
17.8
24.6
23.9
12.3
9.6
2013
12.8
16.6
27
23.5
10.7
2014
11.7
15.2
24.9
25.8
13.6
121. Moreover, as shown in the Table 7.1.1 below, the proportion of households with an annual total
estimated gross value of R&T production/processing between 1,001 and 2,000 GHS increased from
14 % in 2009 to 20 % in 2013, while the proportion with an annual total value between 2,001 and
5,000 GHS increased from 8 % in 2009 to 12 % in 2013, indicating a positive change in R&T-based
livelihoods. The proportion of households with an annual total value of less than 500 GHS on the
contrary decreased from 27 % in 2009 to 21 % in 2013, showing that the very low-profit R&T
businesses are in decline.
Fig 7.1.3: % distribution of the range of total R&T production & processing value from 2009 to 2014
30
25
20
15
10
5
0
0-500
501-1000
1001-2000
2001-5000
>5001
2009
27
20.1
14.1
7.6
26.9
2013
21.4
21.3
20.2
12.1
22
2014
17.8
23.3
22
12.7
20.4
122. These figures show that 15 % of the households crossed the USD 2/day threshold, but only 1 %
moved to higher income levels above USD 4-5/day. Similarly 10 % moved out of low-value R&T
livelihoods of less than USD 1/day value, and 4 % of these made it beyond the USD 2/day, but none
of the households increased their total R&T livelihood value above the USD 4-5/day. This indicates a
clear stagnation in the profitability of R&T livelihoods.
7.1.2 Causes
123. R&T livelihood changes did not predominantly affect access to food but rather access to income,
which on its turn more directly impacted on access to food. More households in RTIMP treatment
areas appear to have gained income and raised above the poverty threshold of USD 1-2 per day due
to positive R&T livelihood changes in the period between 2009-2014.
124. The analysis of the micro-narratives that came out of the livelihood focus group discussions (cf.
figure 7.1.3 below), revealed that in 93% of the experiences of negative R&T livelihood change and
48
94% of the experiences of positive R&T livelihood change in past 5 years, access to food was not
predominantly caused by these changes. Instead, farmers and processors experienced positive R&T
livelihood change as having predominantly affected their access to income and education, and
negative R&T livelihood change as having mainly affected their access to income.
Fig 7.1.4: “The main effect of the livelihood change in my experience is...”
access to education
access to education
25%
18%
(60)
(23)
22%
24%
16%
(21)
(58)
38%
19%
6%
(47)
(49)
(9)
(14)
access to income
access to food
Experiences about positive changes (243)
7%
access to income
access to food
Experiences about negative changes (130)
125. Correlation analysis (cf. Figure 7.1.4) showed that increased access to food of households was
strongly and positively yet not directly (p 0.22, sig 0.000) influenced by an increase in their total
value of R&T production/processing, which is a proxy for R&T livelihood change. Furthermore it
shows that the increase in R&T production/processing value had a more direct influence on
household income (p 0.54, sig 0.000), which in turn also influenced access to food (p 0.32, sig 0.000).
Last, while enhanced R&T production/processing had a strong and positive influence on household
income (p 0.24, sig 0.000), its influence on R&T production/processing value has proven to be much
more linear and direct (p 0.48, sig 0.000).
Fig 7.1.5: Correlations between food, income, R&T
livelihood value and R&T Production/Processing
Access to food
p 0.32
p 0.27
Production/
Processing
of R&T
Annual HH
income
p 0.54
p 0.48
Total Value of
R&T Production/
Processing
126. The number of household members who contributed to the household income changed marginally
from 2009 to 2014. The proportion of households in which only one person contributed was 35 % in
the year 2009. In the year 2013 and 2014 this proportion decreased to 32 %. The proportion of
households in which two persons contributed to the total household income, increased from 39 % in
the year 2009 to 40 % in the year 2013 and further increased to 42 % in 2014. The proportion of
households in which three or more household members contributed was 27% in 2009 and has not
changed since then. Hence the rise in low-income levels cannot be attributed to an increase in the
amount of household members that contributed to the household income.
127. Households have invested in their R&T livelihoods in the past 5 years (2009-2014) mostly in the area
of production, which to a certain extent has contributed to an overall increase in R&T value (and
thus income) for all households involved in R&T production and processing. As production volumes
49
increased as more households moved into R&T farming, production slightly shifted to processing.
Investments and profits however remained limited, which rendered livelihood improvements fragile.
128. Half of all the surveyed households (of which 42 % female-headed, 45 % presently categorized as
poor and 24 % as rich) have not invested in any existing or new R&T production in the past 5 years.
Nearly 44 % (almost 79 % of which were female-headed, 20 % presently poor and 42 % rich)
invested in existing R&T production activities. Only 6 % (of which 55 % were female-headed, 30 %
poor and 48 % richer) invested in new R&T production activities.
129. More strikingly, nearly 89 % of the households (almost half of which were female-headed, and 34 %
poor and 32 % rich) did not invest in any new or existing R&T processing activity. Only 4 % (of
which 68% female-headed, and 20 % poor and 49 % rich) have increased existing processing
activities, while 7 % (of which 71 % female-headed, and 36 % poor and 44 % rich) invested in new
processing activities.
130. Investments in R&T farming and access to farming technologies appear to be negatively correlated
with R&T production as a main household livelihood. Thus as more households moved into R&T
farming, investments in production decreased (p -0.31) while also access to technologies decreased
(p -0.24). This is not the case for R&T processing, thus suggesting a shift from production to
processing as production volumes increased. These correlations are quite significant though not
linear, thus interacting with other variables related to markets.
131. Undoubtedly, some of those who invested in R&T production (50 %) or processing (11 %) have
gained profit from these investments, given the strongly correlated increases of households with a
total R&T livelihood value of USD 2-4/day (10 %) and households with a total income of USD 24/day (15 %). The increase of households with USD 2-4/day R&T value, however, remained
relatively low and R&T livelihood values did not raise above USD 4/day. Also the high percentages
of households that did not invest in R&T production (50 %) or processing (89 %) indicate a limited
profitability.
7.2
Changes in R&T livelihoods and its causes
(link O1+O2+O3I2)
7.2.1 Changes
132. More households have gained a higher annual total value of household R&T production and
processing in the RTIMP treatment areas between 2009 and 2014, contributing to a general income
increase.
133. Findings supporting this statement are presented in Paragraphs § 120 and 121.
134. Investments and profits, however, remained limited. The percentage of households that moved up to
higher value (and thus income) levels remained relatively low (10 % above USD 1/day) and turned
zero at USD 4/day. Also the percentages of households that did not invest in R&T production (50 %)
or processing (89 %) in the past 5 years are relatively high, indicating a limited profitability.
135. Findings supporting this statement are presented in Paragraphs § 118.
7.2.2 Causes
50
136. New R&T varieties and technologies introduced by RTIMP (and its predecessor RTIP) initially
caused an increase in the value of raw and processed R&T produce, which contributed to an
improvement in R&T-based livelihoods across the entire country between 2009 and 2013. The initial
Fig 7.2.1: The change in your story was caused by …
Business relations
Business relations
13%
39%
(31)
(55)
15%
18%
(20)
39%
10%
14%
12%
(95)
(25)
(20)
(17)
Production
issues
Processing
issues
Experiences about positive
changes (240 stories)
Production
issues
Processing
issues
Experiences about negative
success of the new R&T varieties and technologies caused an influx of people into R&T farming,
which led to a substantial increase of production volumes and triggered a spill-over increase in
processing volumes. Inadequate business and market linkages, however, have turned the tide and
caused prices to drop, hence negatively affecting farmers’ and processors’ livelihoods from 2013.
137. Pattern analysis of the livelihood change micro-narratives (cf. Figure 7.2.1below) revealed that in
39 % of the 240 experiences of positive livelihood change, this was mainly due to R&T production,
while in 18 % it was caused by a mix of production, processing and business relations. In 39 % of the
132 experiences of negative change, inadequate business and market relations formed the main cause,
while production and processing had little influence.
138. In nearly half of the 240 stories or experiences of positive livelihood change collected from R&T
farmers and processors, causes were not specified. In the other half, ‘higher yields’ and ‘better prices’
due to ‘new technologies and planting materials’ came out clearly as the main cause. In 70 % of the
132 stories of negative livelihood change, ‘inadequate buyers’ was mentioned as the main cause. In 9
cases it was ‘limited market’ and in 7 a ‘fall in the price of gari’. Box 7.2.1 on the following page
presents an excerpt of some of the most prevalent and remarkable experiences.
Box 7.2.1: Excerpt of farmers’ and processors’ experiences of R&T livelihood changes
Positive change caused by production
Negative change caused by business relations
I now produce enough yams to sell so I make a lot of
money which I can use to take care of my children.
Before the introduction of the new cassava variety,
she only got 2 bags, now she can get 4-5 bags out of
the same size of land, which helps her to feed her
family. (friend’s testimony)
Knowledge on how to plant in lines has enabled me
to get more cassava and more money.
Cultivating yam has given me the money to send my
sister to senior high school and take care of my
I don't have enough money to pay for my children's
school fees because of the limited market
opportunities.
I am unable to properly cater for my children's
educational expenses as well as the needs of my
family. This is because there are no buyers to
purchase my produce.
I am afraid to expand my farm because there are no
buyers.
Presently, I cannot expand my farm because my
last season's produce was not bought to my
51
family.
Because of the higher demand (for cassava and gari)
he can sell more produce and from the increased
income he can take care of his family's health needs
and the household bills. (friend’s testimony)
After the multiplication, he sells the yam seeds and
from the income this generates he can afford to
increase his farming activities. (friend’s testimony)
I farmed the new variety and from the increased
earnings, I re-invested part in maize and used some
to send my children to school.
The planting-in-line method has increased my
cassava yield. With the income gained, I have been
able to purchase a plot of land.
Now she sells all that she farms and from the
earnings she can pay her health insurance and that
of her children. (friend’s testimony)
The yam business has enabled me to build a house,
set up a business in women's clothing for my wife.
My child is also in secondary school.
The training on improved methods of cassava
cultivated has enabled me to improve my yields and
put up a house which my children can take pride in.
The cassava business has helped me to see my three
kids through school (SHS) and the eldest is currently
in the university.
Through the education on the importance of cassava
and supply of planting materials, my father has been
able to support my siblings and myself through
school.
The new variety has lessened the labour required.
Now, I find it easier to plant and get more harvest in
return. Even though I have enlarged my farm size
from 2 acres of land to 3 acres, I still spend less time
on the field.
Because yam is good enough now, we are able to
increase our sales and so make more income. I am
now able to support my husband.
The training and supply of planting materials has
enabled me to further my children's education, and I
also bought a taxi.
expectation.
I have made losses and my children stopped
attending school due to the low price of gari.
My income has reduced, so now I do not have
enough income to buy farm implements such as
cutlasses, hoes, knap sack and a sprayer to work on
my farm.
Because of no buyers, I do not get money to even
buy soap to wash my children's clothes.
The gari I make goes bad because there are no
buyers.
The climate is fairly unpredictable. Preparing the
farm is also expensive. And when I harvest my
produce and try to sell it, people do not purchase.
There are no buyers, and therefore I cannot gain
enough to set some aside to build my house and
cater for my family's needs.
We have good harvest because of the training by
MOFA but the price of cassava is low so we do not
make enough money from selling.
I can't expand my barbering shop since I didn't get
buyers for my produce.
Due to the decrease in the sales of gari, everything
that she makes is just sitting around, so she doesn’t
get any money. (friend’s testimony)
I used to be able to sell all my produce in three
weeks. But now, even after a month, I am unable to
sell. This has drastically reduced my income.
I run a loss when I do not get market for my
produce because I hire land to cultivate my
produce. Thus many times, I do not have money to
pay my children's school fees and cater for my
household bills.
There are no buyers. As a result, I cannot save
enough money to buy the agro chemicals I need for
my farm.
Due to buyers who are cheating the farmers, I am
experiencing low income and unable to pay my
grand children's school fees.
The problem of limited market has made me unable
to pay my family's hospital bills.
139. Especially yam farmers in the Northern zone indicated that positive change was mainly caused by
improved production. Also in the Central Zone, respondents indicated improved production as the
main cause of positive livelihood change, and business relations as well as processing issues as the
main cause of negative livelihood change. In the Gari chains, 31% of the people found that improved
production was the main cause of positive livelihood change, though 38% also indicated a mix of
improvements in production, processing and business relations as the main cause. In nine of the
thirteen cases of HQCF, improved production was sees as the main cause of positive livelihood
change. The small amount of negative changes in the fresh yam chains were mainly caused by
business relations (12 out of 17). Last, in the Southern Zone, the pattern for gari was similar to the
overall pattern for both positive and negative livelihood changes, though with a bit more contrast on
the negative side. While 46% of the respondents found business relations the main cause of negative
change, 27% of the respondents considered production issues as the most important cause of positive
52
livelihood change. Also for the HQCF (only 9 experiences), business relations were seen as the main
cause for negative livelihood change.
140. Aggregated analysis of all the evidence collected on the causal link O1+O2+O3I2 showed that in
13 of the 25 researched districts (52 %), improvements of R&T-based livelihoods as a result of
enhanced production, processing and market-linking in the past 5 years were found moderately
satisfactory (score 4) to satisfactory (score 5), which is strong but not all attributable to RTIMP. 58 Its
strongest contribution was made in the area of production; its weakest in the area of market-linking.
In nearly 70 % of the cases, adoption of the new R&T varieties and technologies and the
enhancement of smallholder production and productivity was found strong and attributable to
RTIMP, while commercialization and market linking was found relatively strong but not entirely
attributable to RTIMP in only 27 % of the cases, and weak and reasonably attributable in the other
73 % of the cases.
141. The other 12 researched districts (48 %) performed moderately unsatisfactory (score 3) to
unsatisfactory (score 2) thus generally weak in terms of impact on R&T-based livelihoods, and in
one case even highly unsatisfactory (score 1) showing no impact and no trace of RTIMP at all.59 In
all these 12 districts, adoption of the new R&T varieties and technologies and enhancement of
production and productivity was still found strong in 46 % of the cases, while commercialization and
market linking was found weak in 100 % of the cases, both mostly inconsistent and thus not entirely
attributable to RTIMP.
142. Generally, where production outcomes appeared stronger and market-linking weaker, also livelihood
improvements were found weaker, indicating a negative effect on livelihoods of excess production in
the absence of sufficient market linking. In Adansi South and Ahafo Ano South (both gari supply
chains), for instance, where enhanced production of cassava was scored higher than 3 (thus strong)
while market-linking lower than 3 (thus weak), livelihood improvements were scored 3 (thus weak).
In these districts, production levels increased due to the introduction of the new varieties and planting
technologies. Initially, this had a positive effect on livelihoods of farmers and processors. High
inflation and insufficient market combined with an unfavourable land tenure system though caused
prices to drop and costs to increase. Problems of market access and transportation due to inadequate
roads exaggerated the problem. Prices offered by GPCs were too low (a bag of cassava was sold in
2014 at 10-13 GHS, while 5 years before it was sold at 35-40 GHS). Hence farmers ran income
losses that negatively impacted on household food and education and discouraged them to continue
growing the new variety and expand their farms. Supply chain facilitation remained absent and no
serious efforts have been undertaken by RTIMP to address the issues in these districts.
143. Positive as well as negative changes in R&T-based livelihoods were in about 32-33 % of the cases
clearly attributable to RTIMP. Apart from the MEF, which by large was not accessible, hence did
not make a noticeable difference, all three other program mechanisms that this evaluation focused on
(DSF, FFF and GPC) were formally in place in most of the cases where livelihood changes were
reasonably attributable to RTIMP. Where livelihood changes were found strong and positive, FFFs
58
Cf. Section 6.1 for the rating system and Table 6.1.1 for the scores. In 3 gari-producing districts (Tano North, Techiman, and
Gomoa East), this contribution was scored satisfactory, thus reasonably attributable to RTIMP). In 7 other Gari supply chains
(Nkwanta South, Damango, Birim Central, Upper West Akim, Mampong and Assin South), 2 fresh yam supply chains (Nanumba
North and East Gonja), and 1 PCF supply chain (Central Gonja), it was found moderately satisfactory, thus inconsistent and not
entirely attributable.
59
In 3 districts producing gari (Suhum, Adansi South and Ahafo Ano South), 1 producing cassava flour for local consumption (Wa
East), and 1 supplying fresh yam for export (Kintampo South), the influence on livelihoods scored moderately unsatisfactory. In 2
gari-producing districts (Agona East and North Dayi/Kpando), 1 yam (Pru), 1 PCF (Wassa Amenfi West) and 2 HQCF, this
contribution was found unsatisfactory. In 1 PCF-producing district (KMA), there was no influence and no improvement.
53
successfully increased smallholder production, while GPCs and other small enterprises actively
contributed to the supply chain linking and commercialization of resource-poor farmers and
processors. Where livelihood changes were rather weak and less positive, insufficient market
linking due to weak DSF performance hampered the growth of farmers’ and processors’ profits and
investments.
144. Analysis of livelihood change narratives (cf. Figure 7.2.2) showed that, in about 33% of farmers’ and
processors’ individual experiences of positive and negative R&T livelihood changes across the 25
researched districts, these changes were predominantly attributed to RTIMP. Farmers and processors
attributed positive changes that occurred in their communities in the past 5 years mainly to RTIMP
(37 %) or to individual efforts (23 %), or a combination of both (16 %). Also negative changes are
highly attributed to RTIMP (28 %) as well as to people’s own efforts (23 %), but partly also to a
combination of RTIMP and other organizations (20 %).
145. Also the aggregated analysis of evidence on the link O1+O2+O3I2 showed a similar figure though
from a different angle thus shedding a different light on the perceived attribution. Instead of
individual experiences of farmers and processors, the configuration analysis took the entire supply
chains (thus the collective experiences of all supply chain actors involved) as the main unit of
analysis, hence bringing a systemic perspective. In 8 of the 25 the supply chains (32 %), the observed
livelihood changes appeared reasonably attributable to RTIMP, which was also confirmed in the
sensemaking workshops in these districts. In 3 of the 8 supply chain areas (Tano North, Techiman
and Gomoa East), the link was scored satisfactory (score 5), meaning strong and attributable to
RTIMP. In 5 of the 8 supply chain areas (Kintampo South, Wa East, Suhum, Adansi South and
Ahafo Ano South), the link was scored moderately unsatisfactory (score 3), meaning weak and
attributable to RTIMP. In 7 of the 8 supply chains, market linking was found weak (score 2-3) mostly
due to weak DSF performance.
Fig 7.2.2: “The livelihood change in my experience was due to…”
other people or organisations
other people or organisations
7%
12%
(18)
(16)
20%
(27)
23%
37%
23%
28%
(57)
(90)
(30)
(37)
16% (40)
my own efforts
RTIMP
(local gov’t)
my own efforts
RTIMP
(local gov’t)
146. Improvements in livelihoods as a result of enhanced production, processing and market-linking was
found the strongest in Tano North, Techiman and Gomoa East (all gari supply chains). The
improvements occurred particularly between 2009 and 2013 due to enhanced production of cassava
and processing of gari, and some market linking, reasonably attributable to RTIMP. In Techiman and
Tano North, all four program mechanisms (DSF, FFF, MEF and GPC) were formally in place, with
the FFFs and GPCs performing very strong and also the DSFs performing better than in most other
54
districts. Gomoa East only had a GPC and DSF, and no FFF or MEF, yet this didn’t make a
difference in terms of livelihood improvements as compared to the other two districts, thus
suggesting their redundancy. The absence of FFF though was compensated by the DSF that had
taken over the training role of the FFF, while MEF was not accessible to small farmers and
processors in almost all researched districts and thus did not demonstrate a noteworthy difference
anywhere. The DSF in Gomoa East was merely used for training purposes and contributed very little
to creating new market linkages and business opportunities, but this was compensated by the GPC,
the Gomoa Obuasi Agrico Cassava Processing Centre, that played a crucial role in helping farmers
and processors to commercialize without much help from the DSF.
147. In the five districts that scored ‘moderately unsatisfactory’ on the link O1+O2+O3I2 (thus weak
but attributable to RTIMP), generally improvements in R&T-based livelihoods remained limited. All
five districts formally had FFFs and DSFs, while the three supply chains that produced gari –namely:
Suhum, Adansi South and Ahafo Ano South– also had functional GPCs. Wa East produced
cassava flour for local consumption and did not have any GPC, while Kintampo South produced
fresh yam for export and therefore did not require any processing centre. Only Adansi South had a
PFI that formally provided MEF funding, but nobody applied and obtained the funding. While FFFs
performed relatively well in these districts (except in Adansi South), DSF performance was found
very weak and its contribution to market-linking virtually nil. High inflation, poor roads and high
transportation costs, low prices for cassava offered by monopolizing GPCs, unfavourable land tenure
practices, and poor marketing has overshadowed the benefits from enhanced production and
negatively affected small farmers’ and processors’ livelihoods and income levels.
148. In 17 of the 25 supply chains, livelihood changes were found not entirely attributable to RTIMP. In
10 of these, they were found relatively strong but inconsistent due to other interventions that
positively affected R&T-based livelihoods and in some cases also due to elite capture of RTIMP
support that excluded farmers and processors from benefitting from the program. In the other 7 cases
(North Dayi/Kpando, Agona East, Pru, Ho Municipal, Tano North/Dua Yaw Nkwanta, Wassa
Amenfi West and Kumasi Metropolitan Assembly), livelihoods improvements were very weak or
virtually absent and were barely or not influenced by RTIMP (score 1-2). As most of the evaluated
RTIMP mechanisms were dysfunctional or not in place in these 7 districts, together they form a
useful comparison group providing counterfactual evidence for the difference that the program has
made in the other 8 districts mentioned above where changes were found reasonably attributable to
RTIMP.
149. In North Dayi/Kpando (gari supply chain), there was no functional GPC, an FFF was held in the
district however there was no FFF in the researched communities, no MEF, and a DSF that had not
been accessible to farmers and processors. Originally, there was a GPC but it was set up purely as a
private enterprise that did not benefit or impact any of the surrounding communities in any possible
way. Although there was no FFF in the sampled communities, production of cassava increased
because more farmers started growing cassava for commercial purpose as the price of cassava went
up. But the available market saturated as all surrounding districts started producing the same gari and
cassava dough, and no new markets could be accessed due to bad roads and the lack of any marketlinking activity. Although there was a DSF in the district, farmers and processors had not been part of
it. Overall, quite strong and consistent evidence (with score 5-6) showed a clear disconnect between
the program and the three researched communities in this district, as community members did not
know of any of the RTIMP interventions.
55
150. In Agona East (also gari), none of the evaluated program mechanisms were there60, but cassava
production increased due to the WAAP that provided free planting materials and weedicides, and the
presence of a starch factory and a GPC that initially provided a ready market for farmers’ produce.
Due to local politics61 and financial defaults, both the GPC and the factory had to shut down.
Meanwhile more than 80 % of the men and women in the Mankrong cluster had turned to cassava
farming and processing as their main livelihood. In the absence of any new market-linking initiative,
the enhanced production negatively impacted on resource-poor farmers’ and processors’ livelihoods
and living standards. Also in this district, the evidence collected was found quite robust (score 6).
151. In Pru as well, none of the program mechanisms were in place. Pru was sampled from the RTIMP
list as a fresh yam district for export and improved yam seed distribution. Yam farmers in Pru,
however, had not seen any seeds of the improved yam varieties, and not heard of any of the
technologies to improve crop husbandry, soil fertility and pest management practices. Before 2009,
they obtained some subsidized weedicides and fertilizers, but these were not from RTIMP. No FFFs
had been organised and no training and starter pack for commercial seed growers to multiply
certified seeds had been provided. There had been no agricultural extension service since 3 years
since extension officers had left the post because of the lack of electricity and other amenities. The
farmers generally have too small fields and, using the old labour and cost intensive planting methods,
don’t gain any profit. Because the farmers don’t have storage barns to preserve their produce, they
are compelled to sell to agents at cheaper prices. The farmers organised as an association that
collectively sells the yam in bulk, but this has not enabled them to bargain for better prices.
Middlemen force them to settle for the cheapest rate, as farmers are usually in need for cash at
harvesting time and don’t have any other market opportunity. Poor roads and high transportation
costs made it impossible for farmers to get their produce for instance to Kintampo market before it
rots and without any damage. The capital-intensive and low-profit nature of yam farming and the
lack of any market-linking services and financial support have prevented about 90% of farmers from
expanding their farms. This has also discouraged young adults from going into yam farming. Less
than 10% of those younger than 35 years are farming yam in Pru. The evidence collected in this
district was found strong but not completely consistent (scores 4-5), because it was more difficult to
cross-check the information since the officials that were interviewed were quite new and people in
the villages didn’t know about RTIMP either.
152. In Ho Municipal and Tano North/ Duayaw Nkwanta (both HQCF), the supply chain leaders,
supposed to be heading the chain, have been ineffective and impact has been limited. In these
districts, the supply chain leaders were private HQCF enterprises but not GPCs. In Tano North, the
supply chain leader stopped operating in 2012, while in Ho Municipal the enterprise merely operated
and purchased from the farmers when it received orders for HQCF, which had been very limited in
recent years. Individual farmers and processors did not process any HQCF and there was no MEF. In
Ho Municipal, an FFF had been held in the district but not in the communities researched. In Tano
North/Dua Yaw Nkwanta, there was an FFF but not on cassava for HQCF, only on cocoyam.
According to the Tano North RTIMP 2nd quarter report (2012), farmers themselves requested an FFF
on cocoyam after a PRA exercise. Regardless, production of cassava in increased tremendously due
to the new varieties and planting methods that had been introduced through the DSF. The DSFs were
60
According to the MoFA officials and the farmers and processors who attended the district sensemaking workshop, Agona East is
merely an RTIMP-participating district, not an implementing district, which should explain the absence of any RTIMP mechanism.
61
The starch factory was brought in by the NPP (New Patriotic Party) when they were in power. The GPC was established by the
NDC (National Democratic Congress). The minister of agriculture at that time was a very high-ranking member of the ruling NDC.
The GPC was set up in his hometown, Mankrong, and the owner of the GPC is his close relative.
56
mainly used for training purposes, not for market linking. At Ho Municipal, buyers of cassava were
not part of the forum.
153. Finally, in the two PCF clusters in Kumasi Metropolitan Assembly, as well as the PCF chain in
Wassa Amenfi West, no trace of RTIMP intervention could be found and no connection between
off-takers (or supply chain leaders) and farmers and processors (or suppliers) identified. The offtakers or plywood industries had never attended any DSF meeting or heard of other RTIMP
initiatives, and expressed no interest in joining the RTIMP, yet they were mentioned in the supply
chain facilitation report dated July 2012. Kumasi Metropolitan Assembly itself did not have farmers
and processors that could supply Bibiani Logging and Lumber Company (BLLC) and Bondplex, the
two supply chain leaders in the district. Bibiani Logging and Lumber Company (BLLC) relied on
cassava flour (kokonte) from two female suppliers in Wenchi in the Brong Ahafo region who
supplied more than double of the 180 bags needed per month. Cassava flour was used merely for
increasing the quantity and thickness of the glue for making plywood, while the glue itself was
supplied from Germany, so not much PCF or kokonte was needed. Bondplex had even stopped using
cassava flour for making the glue about two years earlier, as they found that wheat flour gave a better
result than cassava flour. Wassa Amenfi West on the other hand did have cassava growers and
processors, but Samartex did not purchase from or worked with locals and instead obtained their PCF
from middlemen who bought it in Techiman and Mampong, for which the suppliers could not be
identified. Although the procurement officer of Samartex admitted that transportation costs would
greatly be reduced if the company would acquire the cassava chips directly from Samreboi,
Abekoase and Mpokuase, no relationship between the company and these communities existed.
There were no processing centres in the district, but some individuals did have some processing
equipment that they rented out for a fee, which helped to enhance the production of gari (but not of
PCF for the plywood industry). Production of cassava increased due to the trainings on planting
techniques, the weedicides and the free seedling provided by the WAAP (but not by the RTIMP). No
market linking had been undertaken in this district with support from the RTIMP or the WAAP or
any other program around cassava. Farmers recently shifted from cassava to cocoa farming because
they received a fixed price for the cocoa. Hence farmlands transformed into cocoa plantations.
Changes in the livelihood of the farmers and the processors were found Confirmed by both the
farmers in the FGDs (women and men) and the officials and off-takers in the KIIs, the evidence for
this district was found quite strong and consistent (scores 5-6).
8
Findings on R&T market-related changes and causes
154. In this section, key findings from the analysis of the evidence on the changes and causes with regard
to the market linking are presented. R&T supply chains supported by the RTIMP were expected to
tap into old and new markets and develop into sustainable and inclusive commodity chains.
155. Two assumptions underneath this contribution claim needed to be examined. The first was that
District Stakeholder Forums (DSFs) supported by supply chain facilitators and market-linking
initiatives would help develop sustainable and inclusive R&T commodity chains. Second, more
recourse-poor R&T farmers and processors (including women and young adults) would
commercialize and become part of the supply chains if they would obtain the knowledge and
capacity to increase their production, access markets and develop viable businesses.
156. Two evaluation questions needed to be answered: (a) to what extent these assumptions held true (or
not); and (b) what the major barriers were for farmers and processors to commercialize and access
57
old and new markets. In addition, insights were sought for the learning question: what is needed to
makes the DSF an effective mechanism for business- and market-linking.
157. To answer these questions, the following links in the ToC
needed to be examined:


M1c
M1b
C1a+(M1)O1 which is about positive changes in
market-linking, and the sustainability and inclusiveness
M1a
of these changes, resulting from enhanced capacity in
business and market planning and the organisation of District
Stakeholder Forums (DSFs);
C1b
O1
C1a
M1
M1c+M1b+O3+O2 +O1C1b which is about the commercialisation and supply chain linking
of an increasing amount of resource-poor R&T farmers and processors as a result of enhanced
production, processing and market-linking.
158. The findings related to these 2 causal links are described and explained in greater detail in the next
two Subections 8.1 and 8.2.
8.1 Changes and causes of enhanced market-linking
(link C1a+M1O1)
8.1.1 Changes
159. Market linking of supply chains through the DSF was found weak and ineffective in more than 84 %
of the researched districts across the country. As a consequence, market insufficiency has offset the
initial gains from enhanced R&T production and hampered the growth of farmers’ and processors’
profits and investments. In only 16 % of the researched supply chains, market linking was slightly
more effective yet unsustainable.
160. The aggregated analysis of the evidence collected on the causal link C1a+(M1)O1 showed that in
21 of the 25 supply chains (84 %), market-linking as a result of the DSF functioning and the services
and trainings provided through RTIMP was found quite weak (≤ score 3). Inadequate market linking
combined with overproduction caused market saturation. As prices dropped and excess produce
could not be sold, farmers’ and processors’ profits and investments stagnated, rendering
improvements in their livelihoods fragile.
161. In 4 of the 25 supply chains (12 %, all gari), market linking was found stronger but inconsistent, thus
not entirely attributable to RTIMP (score 4). In these cases, new and more diverse market links were
established that resulted in stronger and more resilient livelihood improvements, yet these have not
proven strong enough to withstand external shocks causing volatile market behaviour and prevent
market saturation.
162. More findings supporting this statement can be found in the former Section 7.2 as well as in the next
Subsection 8.1.2.
58
8.1.2 Causes
163. In 48 % of the researched cases, DSFs to some extent contributed to strengthening the supply
chains, but largely failed to link the supply chains to sufficient markets. Where livelihood
improvements were still found relatively strong, this was due to a stronger performance of other
RTIMP mechanisms as well as interventions of other organisations.
164. In 12 of the 25 researched districts (of which 8 producing gari, 3 fresh yam for export, 1 HQCF, 1
PCF and 1 cassava flour for local consumption), the causal link C1a+(M1)O1 was scored
moderately unsatisfactory (score 3), thus weak but reasonably attributable to RTIMP. Here DSFs
mostly contributed to the strengthening of the supply chain linking between farmers, processors,
collectors, and transporters, but failed to link the supply chains to sufficient markets. In 8 of the 12
districts (6 gari and 2 fresh yam), livelihood improvements were still found relatively strong, which
was largely due to a stronger performance of GPCs and/or FFFs and the influence of other
organisations.
165. In Assin South, where livelihood improvements were scored moderately satisfactory, RTIMP helped
establish a functioning supply chain and a DSF mechanism that revolves around the GPC Dannes
Anointed Ltd. Supply chain linkages were created between the GPC and selective farmers and
processors. DSF meetings with those involved in the chain (average 40 participants) have been
organised since 2012 once or twice a year. The BAC has trained DSF participants on marketing
(including scouting the market before setting the price) and business skills (such as costing, pricing,
packaging and expiration dates).62 The processors at the DSF were merely employees of the GPC and
were taught about packaging and expiration dates for the GPC products. Production of cassava and
processing of gari within the chain increased, and farmers and processors directly involved in the
chain were able to commercialise and improve their livelihoods sustainably. Outside the chain,
however, farmers were struggling to find markets for their cassava as the GPC had a complete
monopoly and was the main price setter in the district. The GPC had a very limited reach with
regards to the number of processors that gained access to its improved equipment and technologies,
and the farmers that supplied the center. The GPC was practically the only processor of gari in the
area. According to the figures from the livelihood analysis, in the past 5 years, the number of women
involved in processing dropped from 17 to 13% and men dropped from 10 to 8%. There was no local
market for gari in Assin South and those involved in processing are mostly employees of the GPC, or
individually produced gari on a subsistence level. The GPC secured linkages for selling gari at higher
demand markets outside the district and export markets in Italy, Canada and the UAE. Hence it filled
a market niche of higher quality gari for export that was not accessible to most farmers and
processors because they simply don’t have the capacity to meet the demands and standards. The
centre moreover had its own farm of 350 acres of cassava to ensure a regular supply of cassava at a
moderate price for its factory. These farms were in existence before the establishment of the GPC but
vastly limited the amount of cassava that the GPC purchased from the farmers in neighboring
communities. Some farmers supplied the GPC but were merely price takers and complained about
the price being too low.
166. Also in Upper West Akim and Mampong (both gari supply chains as well) income levels increased
in particular among farmers but also among processors between 2009 and 2012 due to the successful
introduction of new seed varieties and planting technologies by the FFFs. From 2013 onward,
however, this trend turned negative due to excess production, market saturation and poor pricing.
62
With regards to costing and pricing, since farmers were not in a position to negotiate, the effects of these training must have been
limited, which raises questions around the purpose and relevance of some BAC trainings.
59
DSF meetings were focused merely on strengthening the established supply chain linkages and
creating a platform for the chain actors (farmers, processors and a few buyers) to freely discuss issues
concerning farming, processing and marketing. Through the DSF, the supply chain actors learned
about each other’s needs and built trust. Based on this, an informal credit system was developed63
that addressed the problem of insufficient capital to make the supply chain operational. Through DSF
meetings, farmers64 linking up with various processors, including the GPC, within and outside the
district who came to their farms to purchase the cassava. Processors were also linked to educational
institutions and major buyers outside the district. The main problem for farmers and processors
though was to find new buyers for their produce where the existing markets were largely insufficient.
The majority of supply chain actors interviewed, especially farmers, complained about limited
market opportunities particularly since 2013. This was due to more farmers going into production
without a corresponding increase in market opportunities. Issues of market insufficiency raised at the
DSF were not addressed. Apart from some business trainings organised by the BAC at FFF meetings,
farmers and processors did not receive any business and market linking services. The actual market
linking was mostly done on an individual basis through exchanging contacts. Farmers and processors
were not able to organise their production and trade to reach sufficient scale and quality needed for
entering new markets. In a few occasions where attempts were made to create new market linkages
through the DSF, these were weakened by middlemen interventions (in some cases even MoFA
officials) and by farmers’ and processors’ inability to organise. Hence the amount of participants per
DSF meeting and the amount of meetings held annually decreased, and rarely did major buyers and
traders attend the meetings. Whereas meetings were held 5 to 7 times in 2010 with 50 to 100
participants, in subsequent years it declined to average 1-2 a year with only one third of the original
amount of participants.
167. Also in Nkawnta South (gari), livelihood improvements were found moderately satisfactory or
relatively strong, despite the weak market linking, and this was mainly due to the bigger quantities
and better quality of produce that caused an increase in income levels, which for some even made it
possible to re-invest in businesses and improve the wellbeing of their households. The DSF in this
district performed relatively well in terms of strengthening the supply chain. It enabled the chain
actors to freely discuss and resolve the challenges regarding marketing and pricing without much
interference of the MOFA, and provided all sorts of trainings for the participants. People who
participated in these trainings and attended the DSF meetings obtained higher prices for their produce
because of the higher quality. Average 1-2 DSF meetings were held per year with average 50
participants of which 60-70 % were women and all actors in the supply chain fairly represented
(including farmers, processors, transporters, traders, and GPC leader). The amount of participants has
increased over the years as more farmers, processors and buyers have become aware of the benefits
they can gain from participating in the DSF. New market-linking efforts however have not been very
successful so far. Excessive production of the new cassava variety saturated the existing local
markets. Hence demands and prices dropped, while costs of farm inputs, transportation and
household expenses increased. The price for gari dropped from 3 to 2 GHS per bowl since 2010, or
from 120-130 to 80 GHS per bag, while SHS school fees increased from 200 to 800 GHS. Due to the
increase in transportation costs, buyers from Accra, Mankesim, Bawku, Kasoa and Bolgatanga
ceased coming down to the communities. Attempts to create new market links for cassava chips,
cassava beer and gari for school programs have mostly been unsuccessful. Community members
63
This also happened in many other districts across the entire country.
An RTIMP beneficiary Elvis Opoku, managed to win the price for Ashanti regional best cassava farmer in 2012 and was
connected to the GPC and other farmers through the DSF meetings. He was a commercial seed producer and supplies across
districts.
64
60
started producing cassava chips for Ayensu Starch Factory and for DATCO for producing beer, but
the demands of these buyers were too irregular. District officials mentioned that negotiations were
held with these new industries, but these have not yet materialized. Also the market for cassava chips
to Japan and China has not picked up, as communities lack the necessary equipment and
infrastructure to be competitive. A consultant was brought in to establish value-chains, but not much
has come out of this.
168. Similarly, in West Gonja (gari), efforts to create more sustainable market linkages and ensure supply
meets demand were very limited. The focus was mostly on supplying the existing local gari market.
The DSF meetings, which were attended by 60% male and 40% female participants, provided
opportunities for the different chain actors to establish business connections. This helped to expand
the existing market for gari and cassava from 2009/2010. Through the DSF processors were able to
link up with buyers in distant places such as Burkina, Bolgatanga and Bawku. End of 2014, an order
came in from Guiness Ghana to supply 1,000 tons of cassava per month. Hence RTIMP provided 10
commercial farmers (10 acres each, 50 acres in total) and 6 smaller farmers (5 acres each, 25 acres in
total) with planting materials to speed up the production of cassava. But so far this order hasn’t yet
materialised, and other new market linkages have not been established. Improvements in livelihoods
did occur in one community cluster65 in this district due to the increased yields of the new cassava
variety, the business development trainings from the BAC plus additional technical support provided
by the MOFA and a well performing the GPC in Damongo. In the other communities, livelihood
improvements were rather attributable to various interventions of other organisations (such as a
borehole from World Vision, a processing centre from TechnoServe, training in gari processing from
OIC International etc.). These interventions helped increase the income levels in particular of women,
since the processors were mostly women who also had taken up the farming since the new planting
varieties and technologies had made it easier to produce cassava and obtain higher yields. Market
saturation, changing weather patterns and the invasion of Fulani herdsmen in the area, however,
threatened cassava farming and gari processing, which pushed many farmers into charcoal
production despite the negative impact on their health and environment.
169. In Abura Asebu (HQCF), DSF meetings held 2-3 times a year were mostly used for training
purposes66 and resolved around the GPC Tropical Starch that monopolized the supply chain. Tropical
Starch produced HQCF and starch out of cassava, and was the main buyer of the improved cassava
variety. Farmers complained though about its unreliability and the low price it offered for the cassava.
At times of gluts, Tropical Starch refused to pay the price agreed at the beginning of the season.
Hence farmers preferred not to sell their produce to the GPC. Because of the excess production,
people started processing the cassava into gari. HQCF require highly sophisticated technology and
machinery, and thus is not affordable for individual farmers and processors, so the only other option
was gari. There was also a women’s cooperative in the district, called the Bosomin Group that
produced gari and flour for bread and chips at a small scale. And also employees at Tropical Starch
(all women) produced gari at a small scale. But also in this district, the local gari market was largely
insufficient, and no efforts were made through the DSF to find other markets. Despite the mismatch
between farmers and the GPC in the HQCF supply chain and the weak performance of the DSF in
terms of market linking, farmers’ and processors’ livelihoods improved significantly due to a variety
of programs running concurrently in the district (including RTIMP, WAAP, REP, APFOG, CSIR
65
A community cluster is composed of three communities. In total 30 community clusters were sampled in 25 districts, each
consisting of 3 communities, thus 90 communities in total.
66
Trainings and information provided by the district officials at the DSF meetings focused on seed multiplication, distribution of
cassava sticks, financial recordkeeping, quality assurance, packaging, price-setting and sanitation at processing sites.
61
and other MoFA programs). The issue of market insufficiency though has not been addressed, and
consequently livelihood improvements have stagnated.
170. In Nanumba North and East Gonja, the yam supply chain actors mainly operated within a limited
and existing export market. There has not been any new market linking that is attributable to RTIMP.
Fresh yam for export doesn’t require any processing, so there were no upgraded processing centres or
GPCs that could take the lead in searching for new markets. Although the DSF helped the farmers,
aggregators and transporters to build their relationship and appreciate their interdependence, it hasn’t
helped them find new markets. The lack of new or additional buyers for the increased yam
production has adversely affected the demand for yams, in particular in East Gonja. This decrease in
the demand for fresh yam since 2013 consequently made it impossible for farmers to reinvest and
commercialize. Farmers in both districts confirmed that the prices for yam had generally gone down.
Whereas farmers in Nanumba North could sell a yam in 2011 for 300-400 GHC, they would only
get 120-150 GHC for the same product in 2015. In East Gonja, the effect of the drop in price was
intensified by the increase of fuel prices. In 2013, the transport of 100 tubers to Accra cost 30 GHC
and fetched between 300 - 500 GHC, but in 2014, the transport cost went up to 50 GHC whilst
fetching only 100 -150 GHC. Moreover, in East Gonja, the yam beetle infestation reduced the market
value of the produce. Improvement in the livelihoods of the farmers nevertheless were found
relatively strong in these two districts, as compared to other districts where market linking was weak
and supply chain leaders such as GPCs were absent. The improvements were caused by the enhanced
yam production combined with other influences that are partly attributable to RTIMP and partly due
to other organisations operating in the area. In Nanumba North, Sinapi Aba Trust, ADVANCE and
ACDEP supported farmers with credit and farm inputs. In East Gonja, SIFS and SEND-Ghana
supported farmers with inputs and technologies for crop diversification (maize, rise, soy beans, etc.),
which compensated for the negative impacts of the yam beetle plague. Also activities supporting
farmers of IITA and SNV have been observed in the area.
171. In Suhum, Kintampo South, Wa East and Ho Municipal, where livelihoods were found weak,
other RTIMP mechanisms or interventions from other organisations did not compensate for the weak
performance of the DSF and other negative trends that impacted on resource-poor farmers’ and
processors’ income levels. High inflation, poor roads and high transportation costs, low prices for
cassava offered by monopolizing GPCs, unfavourable land tenure practices, and poor marketing has
outweighed the initial benefits from enhanced production and negatively affected small farmers’ and
processors’ livelihoods and income levels.
172. In 36 % of the researched districts, DSF contribution to developing the supply chains was virtually
nil, which made resource-poor farmers and processors more vulnerable to unfair trade practices
and power abuse by clan leaders and middlemen controlling the farm-gate prices and access to the
local markets.
173. In 9 supply chains (3 PCF, 4 gari, 1 fresh yam and 1 HQCF), the causal link C1a+(M1)O1 was
scored unsatisfactory (score 1-2), which is very weak and barely attributable to RTIMP. Here DSFs
contributed very little to the development of the supply chains and no efforts were made to link
farmers and processors to markets. In 8 of the 9 cases (Adansi South, Ahafo Ano South, North
Dayi/Kpando, Agona East, Pru, Tano North, Wassa Amenfi West, and Kumasi Metropolitan
Assembly), livelihood improvements were found very weak due to weak supply chains67. In most of
these, inadequate DSF performance wasn’t compensated by other RTIMP mechanisms or other
67
Cf. Paragraphs § 147-152 for more details on North Dayi/Kpando, Agona East, Pru, Tano North, Wassa Amenfi West, and
Kumasi Metropolitan Assembly.
62
programs, making resource-poor farmers and processors more vulnerable to unfair trade practices
and power abuse by clan leaders and middlemen who control the farm-gate prices and access to the
local markets. In Adansi South, for instance, the GPC started to compete with small processors for
the local market, due to an increase in operational costs and a decline in buyers for export because of
the increased export taxes, while the increase in processing due to excess production already caused a
saturation of the local market in the absence of any new market creation.
174. In only 16 % of the researched supply chains, supply chain development and attraction of new buyers
was more effective in these three districts than in others, due a stronger performance of DSFs and
GPCs. Yet this was still insufficient to withstand external threads and prevent market saturation.
175. Analysis of the evidence collected on the causal link C1a+(M1)O1 showed that in 4 of the 25
supply chains (all gari), market linking as a result of the DSF functioning and the trainings and
services provided through RTIMP, was found stronger but still inconsistent (score 4). Supply chain
development and attraction of new buyers was more effective in these three districts than in others,
which enabled farmers and processors to expand their businesses. Not only DSFs performed better,
but also GPCs were instrumental in developing the supply chains and linking these to new markets.
Yet these have not proven strong enough to withstand market disturbances, due to the limited
capacity to innovate and expand, further constrained by an inadequate rural infrastructure and
inadequate policy and regulations with regard to fair competition and food quality control and
licencing.
176. In Tano North, DSF performance was among the best and strongly contributed to market linking
and the commercialization of cassava farming and gari processing. DSF meetings and BAC trainings
have played an important role in building a strong supply chain. Farmers, processors, transporters,
buyers and service-providers more directly engaged with each other through the DSF, and solved
their supply and demand issues more effectively than before. Links with new markets such as the
Dua Yaw Nkwanta Prisons, the Sunyani Prisons and the Tepa Market were established through the
GPC Adwenepa with the support from the regional MoFA office and the DSF. Processors from the
prisons learned how to produce 'soya gari' from the GPC and began producing and supplying the
prisons themselves. Adwenepa was also able to carve a niche for its products at the Tepa Market,
which so far has proven to be sustainable. There is currently a section at the Tepa market called
'Adwenepa Cassava Market'. The GPC also came out with innovative products such as “Mix Gari”
that became very popular with SHS schools in and outside the district. In 2012, the GPC was charged
approximately 2,000 GHS for the FDA licensing of the “Mix Gari”, which it was unable to pay.
Hence the production of “Mix Gari” was halted. A shift from cassava to cocoa production from 2013
onwards moreover created a land shortage for cassava production that outweighed many of the
positive effects on livelihoods, which surfaces the issue of land tenure and also shows the
vulnerability of monoculture processing centres in the face of volatile markets.
177. Also in Techiman, evidence shows that even though farmers and processors established new market
linkages through the DSF, these have not proven sustainable. Through the DSF, processors
connected with local shops and supermarkets, and with buyers from Accra, Tamale and even Burkina
Faso. Buyers came from other countries to buy the high quality gari from Techiman. They joined the
chain, attended the DSFs, and started buying regularly from local producers. The DSF has formed a
forum where all the actors involved in the supply chains meet to discuss challenges and share ideas
on possible solutions with regard to supply and demand. The DSF worked well in the sense that it
enabled all the chain actors to see their individual activities as complementary and mutually
beneficial if they succeed to collaborate well. The establishment of 3 GPCs operating as member
associations (instead of enterprises owned by a single person) helped farmers and processors to
63
commercialise and link to new markets. The GPCs developed into market hubs that attracted new
buyers through the better quality of gari and formed an example to small processing enterprises that
burgeoned in the district, operating in a similar way as member associations and creating new job and
income opportunities for more than 400 farmers, processors and many other workers. Excess
production however saturated existing markets leading to low demand and poor prices, while the new
buyers from neighbouring countries ceased coming to Techiman due to the emergence of the Ebola
virus, increasing road tolls, the introduction of border taxes, and an increasing competition as better
quality cassava/gari became available in higher quantities in more places, including some closer to
the border. This was confirmed by all stakeholders at the district sensemaking workshop.
178. In Gomoa East, stronger market linking and livelihood improvements were largely the result of the
strong performance of the GPC. In this district, as in many other districts, the DSF was merely used
for training purposes and contributed very little to creating new market linkages and business
opportunities. But the Gomoa Obuasi Agrico Cassava Processing Centre (the GPC) played a crucial
role in helping farmers and processors to commercialize without much help from the DSF. The
Gomoa Obuasi Agrico Cassava Processing Centre is a cooperative that operates as a profitable and
attractive social enterprise with a wide reach of the surrounding communities. By improving the
quality and packaging of the gari and responding to the feedback of original buyers, the Gomoa
Obuasi Agrico Cassava Processing Centre succeeded in getting its produce sold in supermarkets and
acquiring a new buyer from Techiman. This was not sufficient though to absorb the substantial
increase in cassava and gari production and render livelihood improvements sustainable.
179. In Birim Central, DSF meetings held 3 to 4 times a year created a strong platform for farmers,
processors, transporters, fabricators, marketers and buyers to network and develop their business
relations. As a result, farmers and processors gained access to markets in Gomoa, Swedru, Kasoa,
Oda and Asamankese and established business relations with few buyers from Niger and Nigeria.
Also BAC trainings helped processors and farmers develop skills in record keeping of profit margins
against expenditures. But not everybody had the chance to participate and benefit equally. Only those
who were invited to the DSF meetings also accessed these trainings and services, and consequently
were able to produce and sell more at a lower cost. Farmers and processors who were not invited to
participate in the DSFs suffered from market shortage. Also the GPC Biakoye, although a socialprofit enterprise where members could use the facility and equipment for a fee, was too small to
serve all farmers and processors. For those who did participate, businesses developed between 2009
and 2012 because of an increasing demand for cassava and gari, but from 2013 onward, the demand
decreased and prices dropped due to excess production causing market saturation. While it is clear
that the DSF supported by the BAC had a positive impact on supply chain development and market
linking in this district, and also the GPC contributed to the development of small businesses, their
reach of farmers and processors and their effectiveness in terms of market-linking remained rather
limited due to their limited capacity, which impeded the commercialisation and scaling up of
smallholder production as the basis for sustainable livelihood improvement and rural poverty
reduction. Furthermore, livelihood gains were unsustainable due to market saturation caused by a
lack of product innovation and market creation.
64
180. Substantial inputs, services and technologies have been provided in the area of agricultural
production and agro-processing, but a part of the business trainings provided by the BACs, very
little has been offered or done in the area of market linking and market development for smallholders.
The vast majority of district officials who were in charge of organizing the DSFs, act as if market
linking will take place automatically through the DSF and the GPC.
181. Attempts to create market opportunities and link up with new markets through the DSF occurred in a
few exceptional cases (e.g. Nkwanta South, Tano North, Techiman, Gomoa East, Birim Central). In
more than 75 % of the districts, issues of market linking for resource-poor farmers and processors
were not addressed through the DSF mechanism, and (apart from some business trainings) farmers
and processors did not receive any substantial market linking services. This came out quite strongly
from the aggregated analysis of the evidence on market linking collected from the ‘livelihood
analysis’ FGDs and the KIIs with officials and service providers, for which the average score for
‘strength of evidence’ was nearly 5 (thus strong and consistent).
182. From the KIIs with district officials and RTIMP research leaders, it also came out clearly that there
are insufficient resources and capacities among the officials and service providers at the districts and
the regional level, to conduct proper market analysis and develop plans for integrated VC
development, attract investments for transportation and infrastructure development, promote product
diversification/innovation and support market/demand creation among GPCs and other small
enterprises that have an outreach to the farmers and processors in the surrounding communities,
expand the reach of the DSFs by organising more regular meetings that are open to all supply chain
actors for discussing market opportunities and issues of unfair competition and trade practice, and
finally, undertake appropriate action to address the issues raised at the DSF meetings and propose
legislative and policy change at higher levels needed to make actions at the local levels more
successful.
8.2 Changes and causes of supply chain linking and commercialization (link
M1c+M1b+O1+O2+O3C1b)
8.2.1 Changes
183. R&T production and processing has changed across the entire country from a subsistence (thus
merely food producing) occupation to an income-generating (thus commercial) livelihood activity.
184. Findings from the household survey as presented in Paragraphs § 118-120 showed that more people
have gained more value from enhanced R&T production/processing, and as a result more people
have increased their household income. Findings from the correlation analysis of household survey
data and pattern analysis of micro-narratives, as presented in, further indicated that cassava and yam
have become cash crops in RTIMP areas. The evidence collected in districts where no RTIMP
mechanisms were in place –such as in Agona East, Pru and Wassa Amenfi West (cf. Paragraphs §
148 to 150)– suggests a spill-over of this trend into districts not treated by RTIMP, but also clearly
shows that enhanced production without sufficient supply chain development and market-linking
results in weak and unsustainable livelihood improvements and makes resource-poor farmers and
processors more vulnerable to unfair trade practices (see also Paragraph § 170).
65
185. Commercialisation of resource-poor R&T farmers and processors as a result of enhanced
production, processing and market-linking have remained limited and proven largely unsustainable
in more than 88 % of the researched districts across the country. In only 12 % of the districts, this
was found stronger but inconsistent and not entirely attributable to RTIMP.
186. Aggregated analysis of the evidence collected on the causal link M1c+M1b+O1+O2+O3C1b has
shown that in 22 of the 25 supply chains (88 %), commercialisation of farmers and processors was
limited and weak (score 1-3), which in 9 cases was found fairly attributable to RTIMP (score 3). In
only 3 supply chains (Tano North, Techiman and Gomoa East), commercialisation was found
comparatively stronger.
8.2.2 Causes
187. The main explanation for the weak commercialisation of small R&T farms and processing businesses
is that investments and profits have remained very limited. Market saturation in almost all
researched districts caused the demand and prices to drop, which was largely due to weak and
ineffective market linking combined with overproduction. Rising transportation costs due to high
inflation combined with poor roads and poor market infrastructure moreover prevented buyers
from going to the communities, hence further reducing market opportunities, making small farmers
and processors more vulnerable to unfair competition, and thus limiting farmers’ and processors’
profits and investments.
188. The analysis of the household survey data showed that the percentage of households that had moved
up to higher annual R&T values remained relatively low (10 % above USD 1/day) and turned zero at
USD 4/day. Also the percentages of households that did not invest in R&T production (50 %) or
processing (89 %) in the past 5 years are relatively high, indicating a limited profitability (see also
Paragraph § 130).
189. Due to the introduction of the new varieties and planting technologies, production volumes increased
in more than 76 % of the sampled districts, which positively affected people’s livelihoods between
2009 and 2013 (cf. Paragraphs § 198). High inflation and insufficient markets from 2013 onward,
however, caused prices to drop and costs to increase. Hence investments and profits remained limited,
which rendered livelihood improvements fragile (cf. Paragraphs § 158-160).
190. In 8 of the 22 supply chains (36 %) with weak commercialization, livelihood improvements were still
found relatively strong, which was largely due to an overall stronger performance of RTIMP and
positive contributions of other programs/agencies in the area of production (O2) and processing (O1)
particularly between 2009 and 2013. The results in these districts clearly showed that insufficient
market linking due to weak DSF performance (O3) has overridden the positive effects on
commercialisation (C1b) of the stronger RTIMP performance with regards to production and
processing. More findings supporting this statement can be found in Paragraphs § 158-160.
191. In 11 of the 22 cases (50 %) where commercialisation was very weak and also livelihoods were weak
but reasonably attributable to RTIMP, other RTIMP mechanisms or interventions from other
organisations did not compensate for the weak performance of the DSF, and poor roads and high
transportation costs further limited resource-poor farmers’ and processors’ market opportunities,
rendering them more vulnerable to unfair competition and monopolistic behaviour of GPCs and other
enterprises. More findings supporting this statement can be found in Paragraphs § 162 and 170.
192. Wa East somewhat forms an exception, as there has been an improvement in people’s livelihoods
and wellbeing in the area due to the supply chain development of cassava flour for local consumption,
66
which generated increases in savings and investments in fixed and moveable assets and other
livelihoods activities. The interesting thing about Wa East is that, since there was no GPC, farmers
and processors mostly engaged in drying cassava for processing it into flour to make Kokonte, which
is part of the local staple food. While the production of cassava improved due to the introduction of
the new varieties and technologies, the DSF helped stimulate market activity in the area. The shorter
maturity period of the new cassava variety enabled farmers to harvest twice per year instead of once.
As a result, productivity increased from 4 sacks to 10 sacks of dry cassava per 2 acres, and income
from 50 to 70-90 GH₵/sack in about 3-5 years’ time, which has led to more people joining in cassava
farming. Some processors were trained and taken on a learning tour to the GPC Damongo68 in West
Gonja to learn a few basics about processing gari and cassava flour to improve their old-school
processing practices. In the FGDs, supply chain actors confirmed they benefited from the DSF
activities, the use of the ESOKO platform, and the learning visit to Damongo. But unfair competition
created distrust and disadvantaged resource-poor farmers who were exploited by middlemen and the
so-called “market queens” of Wa. They sell their produce on credit to processors and traders, yet
processors and traders are not willing to pre-finance farm inputs, and transporters are not prepared to
carry farmers’ produce on credit. This clearly suggests the need for a certain level of trust and
collaboration as an important precondition for successful supply chain linking and commercialisation,
for which appropriate regulation and mechanisms need to be in place that must ensure fair
competition. Another insight from this particular case is that the lack of access to improved
processing equipment (e.g. for gari) has helped prevent the market saturation that occurred in almost
all other districts. Production, processing and market linking were clearly more attuned to the limited
capacity and infrastructure available in this district. Hence the need for a more adaptive and
integrated approach of value chain development, in which activities and budgets are planned for each
of the supply chains based on the analysis of market opportunities, infrastructural conditions, and
local capacity to create market value, attract investors and prevent unfair competition.
193. In 12 % of the supply chains (in 3 of the 25 districts), GPCs have proven an important mechanism
(as supply chain leaders) to make it possible for resource-poor farmers and processors (as suppliers)
to develop small profitable businesses and gradually grow and commercialize. The success of this
mechanism in these three cases was largely due to its capacity to innovate and create new market
value/demand, its reach of farmers and processors in the catchment area, and the trust it has built
between the various supply chain actors and their buyers and service providers to establish a strong
supply chain.
194. In the 3 supply chains where commercialisation was found comparatively stronger, this was largely
due to a stronger GPC performance with a wider reach, stronger capacity to create demand, and
stronger supply chain relationships. In these three districts livelihood improvements were
comparatively the strongest of all 25 researched districts, due to an overall stronger performance of
RTIMP. In Gomoa East, the GPC “Gomoa Obuasi Agrico Cassava Processing Centre” played a
crucial role in helping farmers and processors to commercialize without much help from the DSF.
The Gomoa Obuasi Agrico Cassava Processing Centre is a cooperative that operates as a profitable
and attractive social enterprise with a wide reach of the surrounding communities. In Techiman, the
establishment of 3 GPCs operating as member associations helped farmers and processors to
commercialise and link to new markets. The GPCs developed into market hubs that attracted new
buyers also from abroad through the better quality of gari it produced, and formed an example to
small processing enterprises that burgeoned in the district, operating in a similar way as member
68
This GPC hopwever did not act as a supply chain leader for Wa East, and did not purchase any flour from the processors or
cassava from the farmers in Wa East.
67
associations. In Tano North, both the DSF and the GPC played an important role in building a
strong supply chain and also strongly contributed to the market linking and commercialization of
cassava farming and gari processing. Farmers and processors were able to access new markets (such
as the Dua Yaw Nkwanta Prisons, the Sunyani Prisons and the Tepa Market) through the GPC
Adwenepa with the support from the regional MoFA office and the DSF. The GPC also came out
with innovative products such as “Mix Gari” for SHS schools. More details can be found in the
Paragraphs § 173-178.
9
Findings on R&T production-related changes and causes
195. This section presents the findings from the aggregated configuration analysis of the evidence on
R&T production-related changes and causes, which was the second component of the RTIMP. By
introducing improved planting materials and technologies through the FFFs, R&T farmers were
expected to increase their production and productivity to the extent they would start making
sufficient profit to expand, organize and commercialize their farming. It was hoped that this then
would also enable farmers to access business finance either through the MEF or through other
channels.
196. The key assumption of RTIMP with regard to production was that FFFs would enable resource-poor
R&T farmers and seed producers to become commercial growers by organising into FBOs and
adopting improved planting materials and technologies. The main evaluation question was to what
extent this assumption held true and under which conditions. The focus of evaluation was on the
FFF’s reach in particular of illiterate farmers, women and young adults.
197. To answer this question, evidence was collected on two important links in the ToC, namely:


C2a+C2bO2 which is about changes in resource-poor
farmers’ and seed producers’ productivity and the total volume
of R&T production in the supply chains, as a result of changes
in farmers’ adoption of improved planting materials and
technologies and their ability to organise and collectively
access credit and markets;
M2c
C2b
O2
M2b
C2a
M2a
M1c
M2a+M2b+(M2c)C2a which is about changes in resourcepoor farmers’ and seed producers’ access to improved planting
materials and technologies due to the FFFs and the development and distribution of bio-agents.
two Subsections 9.1 and 9.2.
9.1 Changes and causes of enhanced R&T production
(link C2a+C2bO2 in the ToC)
9.1.1 Changes
199. R&T Farmers’ and seed growers’ productivity and the total volumes of R&T production have
increased substantially in about 76 % of the researched supply chains, as a result of the adoption of
improved planting materials and technologies. This is mostly attributable to RTIMP. The substantial
increase though has created gluts that caused a saturation of local markets, which hampered farmers’
68
profits and investments and their ability to further commercialise. In 24 % of the researched supply
chains, production/productivity increases were found weak, which was mostly attributable to RTIMP.
200. The configuration analysis of the evidence collected from the livelihood analysis FGDs with farmers
and KIIs with district officials on the causal link C2a+C2bO2 showed that in 19 of the 25 districts
(76 %), increase in production and productivity as a result of R&T farmers’ access to and adoption of
improved R&T seed varieties, technologies and inputs for improving crop husbandry, soil fertility
and pest management practices was fairly strong. In 11 of these districts, this positive outcome was
fairly attributable to RTIMP (score 5). In the other 8 cases, not only RTIMP but also other programs
such as WAAP contributed to the outcome. In 6 of the 25 districts (24 %), production and
productivity increases were found weak.
201. Findings regarding market saturation and the stagnation in farmers’ profits’ and investments’ are
presented in Subsections 7.1 and 8.1.
9.1.2 Causes
202. Were R&T production and productivity increases were strong, this was due to the widespread
adoption of the improved planting materials and technologies. Where the R&T production and
productivity increases were weak, this was due to the limited adoption of the new technologies and
other influences (such as beetle infestation, changing weather patterns, limited markets, the land
tenure system, affordability of inputs, and in 2 cases a total absence of any farming or extension
activity).
203. In the 11 cases (Techiman, Gomoa East, Tano North, Assin South, Birim Central, Upper West Akim,
Mampong, Abura Asebu, Nanumba North, Wa East and Ho Municipal) where increase in production
and productivity was scored strong and consistent (score 5), a high degree of adoption of the
improved planting materials and technologies (such as planting in lines using appropriate spacing and
the application of agro-chemicals) was found. Increasingly more people were entering into R&T
farming using the new seeds and technologies. In the 8 cases (Nkwanta South, West Gonja, Central
Gonja, Suhum, Ahafo Ano South, Kintampo South, Agona East and Wassa Amenfi West) where
enhanced production and productivity was strong but inconsistent (score 4), this was caused by either
a limited adoption of new planting materials, or a widespread adoption due to other programs
operating in the districts (e.g. WAAP). In some of these also unreliable weather patterns and
destruction of farms by Fulani cattle had occurred.
204. In 6 of the 25 districts (24 %), production and productivity increases were found weak. In 4 of these
(East Gonja, Adansi South, North Dayi/Kpando, and Tano North/ Dua Yaw Nkwanta) this was due
to the limited adoption of the new technologies, which was found attributable to RTIMP (score 3).
Also in East Gonja the yam beetle infestation affected the yields. In Adansi South, R&T markets
saturated and people could fetch a quicker and higher income with gold digging than with cassava
farming, which resulted in more than 80% of the community members (especially the youth) in the
supply chain area shifting to gold mining. In Tano North/Dua Yaw Nkwanta, there was an initial
increase in production due to promises of a ready market but subsequently a drastic decrease when
this market did not materialize. In the 2 cases where enhanced production and productivity was
scored very weak with no trace of RTIMP, there was a total lack of access to improved planting
varieties and technologies (e.g. Pru) and access to farmland (e.g. Kumasi metropolitan Assembly). In
Pru (yam supply chain), the absence of extension services, the capital intensive nature of yam
farming, and the lack of loan facility for yam farmers have prevented about 90% of farmers from
69
expanding their farms. In Kumasi Metropolitan Assembly, there were no farmland, only a PCF offtaker with no traceable links with its suppliers.
205. According to the FFF research leaders, a major issue for the sustainability of the adoption of the new
varieties and technologies though is the land tenure system in Ghana. People don’t own land, so the
following year after they have received the inputs and technologies through the FFF and used them
on the land assigned or lent to them, they have to move to another plot and all they have achieved is
lost. Another issue is the affordability of the inputs and technologies particularly for resource-poor
farmers once they graduate from the FFF.
206. Various informal and formal organisations of farmers have been recorded across the country. These
farmer-based organizations (FBOs) helped RTIMP and other government and donor programs better
target and reach farmers, yet so far have proven insufficient to protect farmers against unfair
competition and help them obtain business finance, find new markets and commercialize their
production.
207. Generally, there has been a growing interest in organizing farmers into FBOs, as it is found an
effective mechanism for better reaching farmers with extension and other services, and because of
the alleged advantages in terms of increased bargaining power, credit-worthiness, an d collective
power to influence policies. In the KIIs with FFF research leaders, it was stated that FFFs have been
encouraged to establish FBOs with a leader, secretary and a collective account, following the “Meda”
model (cf Textbox 9.1.1).
208. A study on the efficiency and effectiveness of FFFs under RTIMP conducted in 2014, revealed that
farmers’ membership of formal or informal farmer-based organizations (FBOs) across the country
was quite high, in particular among FFF participants69. Farmers who did not belong to some sort of
FBO were mostly non-FFF participants (78 %), suggesting an influence of farmer organisation on
FFF participation. The study further stated that most farmers either maintained their FFF group or
converted their group into registered FBOs. (MoFA, 2014b) Although in multiple occasions FFF
participants had expressed the desire to maintain their groups, the PIALA evaluation team did not
come across sufficient evidence that could confirm this. The IFAD supervision mission report of
November 2014 stated that FBOs had successfully been developed out of FFFs under RTIMP in 4
locations, but none of these groups had yet been able to obtain credit or bargain better market prices
(Macpherson, Nov 2014).
209. FBOs have been recorded by the PIALA evaluation team in
9 of the 25 researched districts (36 %), of which 2 in the
North (Nanumba North and Central Gonja), 3 in the Centre
(Tano North, Pru and Kintampo South), and 4 in the South
(Assin South, Suhum, Gomoa East and Abura Asebu
Kwamankese). None of these groups stemmed from FFFs or
were influenced by RTIMP. In only 2 of these cases were
farmers able to access business finance (Nanumba North and
Tano North).
210. In Assin South, neither the FBO nor its members were able
to access credit since its inception due to a lack of capacity.
In Gomoa East the FBO called Agrico, which is a mixed
Textbox 9.1.1: From the KIIs with
the research leaders:
To avoid that the owners of land for
the test fields prevent other farmers
to join and participate, the farmers
in the FFF are encouraged to
establish an FBO with a leader,
secretary and a collective account,
following the “Meda” model. Gold
digging has depleted the soil in
many areas. Due to the gold
digging, many farmers have lost
their farms. Cassava can help
restore soil fertility. “Meda” is used
to help the farming communities to
reclaim abandoned lands.
69
According to the figures in the report of the 2014 study on FFFs: approx. 83 % of the FFF participants compared to 37 % of the
non-FFF participants in the central area of Ghana; 100 % of the FFF participants compared to 82 % of the non-FFF participants in
the North; and 91 % of the FFF participants compared to 58 % of the non-FFF participants in the South.
70
group of farmers and processors established before the start of RTIMP but subsequently supported by
RTIMP, was unable access credit and bargain better market prices because of market saturation. In
the two yam supply chains (Pru and Kintampo South) farmers associations were formed to collect the
yam for obtaining the bulk amount required by the exporters in Accra. The associations had limited
bargaining power due to the limited market opportunities.
211. In the other 16 districts where no FBOs were found, farmers were unwilling or not interested in
forming groups due to a lack of trust among farmers and limited benefits. The price of the
commodities was not determined by the farmers and farmers knew they would not get any credit so
forming a group was found useless. There is no evidence that indicates if and under which conditions
FBOs could be an effective mechanism for helping farmers bargain better prices, fight unfair
competition, obtain business finance, access markets and commercialize. All evidence points to the
need for more market opportunities (and thus better roads and market places, and policies and
regulations more supportive of smallholder business development) to enable farmers to
commercialise in the first place, and to the need for developing mixed agri-business organisations
that are less centred around merely farming and more attuned to value creation (thus including agriprocessing and market-linking activities). In Tano North for instance, the FBO at Adwenepa GPC
that also formed an FFF group under the RTIMP, was able to bargain better market prices for gari at
the Tepa market, because their brand/produce was well known for its high quality. They also
obtained finance granted by the district assembly to buy tractors. Here is an example of an FBO that
works because there is market, support and capacity.
9.2 Changes and causes of access to planting materials & technologies
(link M2a+M2b+M2cC2a)
9.2.1 Changes
212. FFFs have effectively promoted the adoption of new R&T planting technologies and seed varieties
in 84 % of the researched districts (21 of 25) across the country.
213. Aggregated analysis of the evidence collected from the FGDs and the district-level and national KII
showed that in 14 of the 25 researched districts, adoption of new planting technologies and new R&T
seed varieties was very high and fairly attributable to RTIMP (score 5-6). In all these cases, FFFs
were mostly effective. In 7 other districts adoption of improved varieties and technologies was scored
relatively strong but not entirely attributable to RTIMP (score 4). In 4 of these, FFFs were most
effective, while in the other 3 districts (Agona East, Wa East and Wassa Amenfi West) high adoption
rates were the work of other programs such as WAAP. Only in 4 districts was the adoption of the
new varieties and technologies found weak, one of which was due to RTIMP (score 3 –namely: Tano
North / Dua Yaw Nkwanta) and the other 3 with barely any RTIMP influence at all (score 1-2 –
namely: North Dayi/Kpando, Pru, and Kumasi Metroplitan Assembly).
9.2.2 Causes
214. Adoption successes were mainly due to the unsurpassed efficacy of the planting in rows using
appropriate distances and agrochemical application technologies, and the visible benefits in terms of
a substantial increase in quantity/quality and value (in particular for cassava). FFFs formed an
effective mechanism to demonstrate and convince farmers of the advantages of the new varieties and
technologies, because of its highly participatory character.
71
215. In the 14 cases (Tano North- Apesika, Techiman, Gomoa East, Assin South, Birim Central, Nkwanta
South, Upper West Akim, Mampong, West Gonja, Abura Asebu Kwamankese, Nanumba North,
Central Gonja, Ahafo Ano South, Ho Municipal) where the rate of adoption of planting materials and
technologies was scored more than satisfactory (score 5-6), this high rate of adoption was due to
access to /free distribution of new planting materials and proven effectiveness of technologies shown
by open days (where whole communities were invited and shown the difference between new variety
and old crops planted). The new planting materials had shorter maturity periods which enabled
farmers to harvest produce twice in a year. The new materials and techniques were shown to increase
yield. Due to the significant increases in R&T yields and value and thus the visible effects of using
the new varieties and technologies, other farmers in the communities of the supply chain area started
to ask FFF beneficiaries to teach them about the new farming, thus causing an influx into the new
R&T farming.
216. According to the FFF research leaders FFFs were so effective because of the way they operate:
“In FFFs everybody learns from each other, instead of the scientists and extension officers merely
teaching the farmers. To establish an FFF, land is made available either by the community chief or
an individual with sufficient land. On this piece of land, 3 types of plots are created: (1) a PAR plot
(Participatory Action Research) on which the farmers, scientists and extension agents conduct
research together; (2) an ICM plot (Integrated Crop Management) on which improved agricultural
practices are tested and demonstrated; and (3) a FPP plot (Farmer Practice Plot) on which farmers
experiment and practice what they learned. All the farmers who participate in the FFFs go through a
full cycle consisting of 5-6 sessions during one year and 4 months. Originally it was one year, but
that appeared not enough as the tubers were still too thin, hence 4 months were added. Through
participating in in a full cycle, farmers learned all the technologies and adopted them on their own
land.” (quoted from the KII notes with FFF research leaders).
217. Although in many communities, women are much more involved in cassava production then men, and
women traditionally do most of the work in R&T farming, FFFs generally targeted and reached
small-scale male farmers between 40 and 60 years old who own a bit of land (2 ha). Since R&T have
turned from merely a food crop into a cash crop, men took a greater interest in growing R&T, and
FFFs have encouraged and supported this. In the cassava supply chains, FFFs have insufficiently
reached women and young adults (<35 years old). In the yam supply chains, FFFs have been more
successful at reaching young adults but less successful at reaching women.
218. The FFF research leaders acknowledged that the FFF target group consisted mainly of small-scale
male farmers with a bit of land (max 2 ha) who are interested in applying the new varieties and
technologies, are willing to allow other farmers access the test fields, and are accessible to the
researchers (thus living not too and close enough to a road so that the multi-disciplinary teams can
get there). The original idea was to reach out to the poorest farmers, but these don’t have sufficient
starting capital (resources and assets) and are too risk averse to become commercial farmers. They
are typically illiterate, sustenance farmers with little or no land (max ½ ha), limited assets and
transportation means, limited access to inputs, and highly vulnerable to shocks such as drought or
unstable prices, with limited bargaining power as they lack the means to get to the markets or find
buyers. The use of mobile phones has helped them, but they remain very vulnerable and are
systematically cheated by the middlemen and small enterprises. Their lack of credit is their biggest
problem. According to the FFF research leaders, only a few resource-poor farmers have participated,
though these formed a very good example for other resource-poor farmers, showing that it is possible
for them to apply some of the technologies that are affordable and gain profit out of it.
72
219. The 2014 study on FFFs found that male farmers form the largest number of the FFF beneficiaries
across all the RTIMP zones. About 66 % of the FFF beneficiaries interviewed in the 2014 study were
male, and 34 % were female. (MoFA, 2014b) These figures are along the same line as the ones
obtained from the FFF constituent feedback sessions and the KIIs in the present impact evaluation.
The FFF constituent feedback responses showed that in the cassava supply chains (gari, HQCF, PCF),
approximately 40% of the farmers who directly benefitted from the FFFs were women. In the yam
supply chains, only 30% of all the FFF beneficiaries were women. If for women, merely young
unemployed women participated, or married women who had their own land and were very active.
While young adults made up only 37% of FFF participants in cassava supply chains, in yam supply
chains they consisted of 57% of FFF participants.
220. Also the FFF research leaders confirmed that, although the women are the cassava growers and not
the men, FFFs mostly targeted and reached the household heads, which in most cases were men
between 40 and 60 years old. Men only do the land preparation, while women do all the rest of the
work. Cassava has always been the women’s sustenance crop used for feeding the household.
However, where cassava has become profitable, men took over and FFFs encouraged this. Also
young people have shown greater interest in past years, as their perception has changed about
cassava from being a women’s and elder’s crop to a cash crop, but the dropping market prices in
recent years have demotivated them.
221. The majority of FFF beneficiaries stated that they had been able to apply what they learned at the
FFFs, which helped them expand their businesses. Young farmers (< 25 years) and women were
generally less positive and more negative and felt less confident to express their needs and ask for
help at FFF meetings than adult men.
222. The constituent feedback graph below (see Figure 9.1.1) shows that average 74 % of the FFF
beneficiaries found that the FFFs had helped them expand their businesses (score 4-5), while 20 %
were neutral (score 3), and 6 % were negative (score 1-2). Women (12 % ) and young farmers < 25
years (14 %) were more negative than the men between 40 and 60 years old (3 % ) while also being
less positive (69 % and 45 % respectively, compared to 77 % men between 40 and 60). The figures
of the national average are almost identical to the ones of men between 40 and 60- years old.
Fig. 9.1.1
223. Moreover, average 66 % of the FFF beneficiaries stated they had been able to apply what they
learned (score 4-5), while 25 % were neutral (score 3), and 9 % were negative (score 1-2). Again
women (12 %) and young farmers < 25 years (29 %) were more negative than men between 40 and
60 years old (3 % ) while also slightly being less positive (62 % and 29 % respectively compared to
68 % men between 40 and 60). (see Figure 9.1.2)
73
Fig. 9.1.2
224. According to the FFF research leaders, women should have been more separated from the men in
women-specific FFFs to obtain better results. In most tribes, women don’t talk or participate in the
FFFs. The constituent feedback analysis indeed showed that women and farmers younger than 25
years felt less confident to express their needs and ask for help at FFF meetings than the average
male participants between 40 and 60 years (14-15 % women and young farmers compared to 4 %
adult male farmers).
10 Findings on R&T processing-related changes and causes
225. This section describes the main findings from the aggregated configuration analysis of all the
evidence collected on the causal links of the contribution claim on processing, which was RTIMP’s
third and last component. Here it was expected that by teaching farmers and processors about
improved processing technologies at GPCs, training them in business development, and creating the
possibility for them to obtain credit combined with grant funding, they would invest in improved
equipment and technologies and expand their businesses, which would lead to greater volumes of
processed R&T of higher quality that is suitable for new markets.
226. The assumptions underneath this claim were two-fold: (a) well-trained processors and farmers would
be able to obtain a loan through the MEF to invest in their businesses; and (b) GPCs would reach and
teach resource-poor farmers and processors about good quality processing practices and the use of
improved technologies and standardized equipment, hence helping them access the MEF and develop
profitable businesses.70
227. The main evaluation question was to what extent the assumptions held true. The focus of evaluation
was on the reach of the GPCs, the accessibility of the MEF, and the contribution of these two
mechanisms to the development of smallholder businesses and commercialisation. In addition, the
evaluation also sought to obtain insights with regard to one key learning question: What conditions
need to be in place for GPCs to become profitable and attractive businesses particularly for young
adults living in remote areas? What supports or hinders GPCs to better link the supply chain farmers
to old and new markets, and how is this influenced by the DSF?
70
Although all assumptions were extensively discussed, reviewed and approved at the evaluation design workshop (before
fieldwork began), RTIMP officials who had participated in this workshop explained at the national sensemaking workshop (after
the field work was finished) that in general it was not the responsibility of the GPC to reach and teach farmers and processors.
Amendments to the assumptions however should have been made in the design workshop. Moreover, IFAD funding targets rural
poverty by enhancing small farmers’ and processors’ ability to develop businesses and access markets, and thus one would expect
that the funding for the upgrading of enterprises into GPCs should contribute one way or another to the development of these small
businesses by exposing farmers and processors to good practices and providing them with access to improved technologies and
equipment. The extent to which this was realised has been inquired by this evaluation as ‘reach’.
74
228. To answer these questions, evidence was collected on the following two causal links in the
contribution claim on enhanced processing in the Theory of Change:


(M3b)+C3cC3bO3 which is about changes in the total
processing volume and quality of the supply chains, as a
result of changes in processors’ ability to develop profitable
businesses due to changes in access to business finance and
market-linking services, and training in good quality
management and the use of standardized technology and
equipment at the GPCs;
M3c
C3c
M3b
C3b
M3a
C3a
O3
(M3c)+C1a+M3bC3c which essentially is about changes in access to business finance and
market linking services, due to changes in farmers’ and processors’ capacity to develop viable
business plan and use standardized technology and equipment taught at the GPCs.
two Subsections 10.1 and 10.2.
10.1 Changes and causes of enhanced R&T processing
(link M3b+C3cC3bO3)
10.1.1 Changes
230. Processed volumes of cassava increased considerably in about 50 % of the supply chains as a result
of more people processing cassava and expanding their businesses by gaining access to training and
facilities at GPCs. In the other 50 % of the cassava supply chains, enhanced processing was found
rather weak.
231. The configuration analysis of the evidence collected from the FGDs with farmers and processors and
the KIIs with district officials and GPC leaders showed that in 9 of the 18 supply chains that involved
processing activities71 (50 %), enhanced processing as a result of more people growing and
processing cassava at a higher quality was found quite strong. In only 3 of these cases (Tano North,
Techiman and Gomoa East), RTIMP contribution was scored more than satisfactory (score 5-6),
which means quite robust and consistent. In 6 of these cases (Assin South, Nkwanta South, Upper
West Akim, Mampong, Suhum and Adansi South) RTIMP contribution was found moderately
satisfactory (score 4), which is strong but rather unclear and inconsistent or not entirely attributable.
232. In the other 9 of the 18 cassava supply chains (44 %), the link (M3b)+C3cC3bO3 was scored
weak. In 4 of these (Birim Central, West Gonja, Abura Asebu and Agona East) the weak results were
fairly consistent and attributable to RTIMP, while in the remaining 5 (Tano North/Dua Yaw Nkwanta,
Ho Municipal, North Dayi/Kpando, Wa East and Ahafo Ano South) RTIMP had little or no influence.
10.1.2 Causes
233. In the 3 cases (Tano North, Techiman and Gomoa East) where enhanced processing was found fairly
robust and attributable to RTIMP, GPCs have proven an important mechanism (as supply chain
leaders) for making it possible for resource-poor farmers and processors (as suppliers) to develop
small profitable businesses and gradually grow and commercialize. Their success was largely due to
their comparatively stronger capacity to innovate and create new market value/demand, their
71
Only gari and cassava flour require processing, thus only 18 of the 25 sampled districts had processing activities in the inquired
supply chains (15 gari, 1 cassava flour for local consumption, and 2 HQCF).
75
ability to reach a larger amount of farmers and processors, and the trust they built between the
various supply chain actors and their buyers and service providers, essential for developing strong
supply chains.
234. The distinct feature in these three highest scoring districts was that the GPCs operated as social profit
enterprises: the Adwenepa GPC in Tano North, the Agrico Cassava Processing Center in Gomoa
East, and the Hansua Womens Group in Techiman. In Techiman, there was also another GPC,
Aworowa Cassava Processing, that was privately owned and operated as a private enterprise but
where community members were permitted and encouraged to use the facilities. These GPCs had a
wide influence and reach in the supply chain areas and contributed significantly to the development
of small farmers’ and processors’ businesses and livelihoods.
235. In the 6 cases (Assin South, Nkwanta South, Upper West Akim, Mampong, Suhum and Adansi South)
where enhanced processing was found strong but inconsistent (score 4), this was caused by GPC
operations with a more limited reach on the one hand, and by the spill-over of excess production
into processing72 that used both new and traditional equipment on the other. Farmers started to
process their excess cassava into gari but mostly in the old fashioned way, yet were able to produce
more and better quality gari due to the new cassava variety. Hence the adoption and/or use of
improved processing technologies and standardized equipment by resource-poor farmers and
processors remained rather limited.
236. In Suhum, Upper West Akim and Mampong, processed volumes have increased considerably but
mostly due to the excess production of the improved variety of cassava, and although influenced by
the demonstrations and trainings at the GPC, mostly by using traditional processing equipment
and technologies. Farmers did not get buyers for cassava, so they started to process it into gari, since
gari lasts and sells better than cassava. Processors were making more and better quality gari due to
the new cassava varieties, causing more people to invest into gari production.
237. Suhum experienced a strong increase in production of cassava that also triggered an increase in
processing of gari, partly due to the GPC Charity & Co which was a functioning individually owned
GPC, but mostly due to the surplus of cassava in the absence of buyers at farm gate. The number of
processors grew by about 5% in past years, while processors also enhanced their volumes. The GPC
had its own cassava farm and purchased very little from the farmers in and around Amanase, and
only when there were plenty of orders. The GPC didn’t appear to run short in buyers though: it
started with 1 and at the moment of the evaluation now had 5 buyers. But farmers also complained
about the price that was paid by the GPC as being too low. The GPC organised demo’s and provided
training only for its employees (27 people). The GPC leader explained that the GPC facilities were
open to other processors who are not employees to use but processors in the neighbouring
communities (mostly women) did not seize this opportunity and were largely unaware of this.
238. Upper West Akim, gari production increased mainly due to excess cassava production, and not so
much as a result of the operations of the GPC Jenefal farms. The GPC did not have the resources to
purchase from the farmers, develop and expand as a profitable business, and build a strong supply
chain. Due to the lack of market and reasonable prices for cassava, farmers began processing their
own cassava into gari. The GPC was open to community members for demonstrations and to use the
72
Also the correlation analysis of the household survey data clearly indicated such a shift from production to processing as a result
of excess production (cf. Paragraph § 126 in Section 7.1.2).
76
equipment for a small fee, yet people were largely unaware of this. Most processors in the
communities (mostly women) still used their old methods and equipment to process gari.
239. In Mangpong, enhanced processing of high quality gari was observed at the GPC Josma and among
processors in the Krobo community, but not in the other communities in the sampled cluster. The
GPC had an impact mostly on the livelihoods of processors in Krobo, in particular among those who
had attended the demonstrations at the GPC. By using improved technology and equipment (such as
grinding machines), the processing of cassava became more time-efficient and profitable.
Particularly in Krobo, youth increasingly showed an interest in cassava processing. However, apart
from those women who had participated in the demonstrations, the GPC did not lead to more
processors in the district having access to improved processing equipment. Although they
appreciated the upgraded processing equipment used at the GPC, most processors could not afford
them due a lack of finance. Many processors in Krobo still used old equipment even though they had
been trained at the GPC.
240. In Assin South and Adansi South, processed volumes and quality increased due to the use of
improved technologies and equipment by the GPC, but with limited reach and benefits for
resource-poor farmers and processors. While in Assin South people did not trust and did not have
access to the GPC that monopolised the market, in Adansi South their access to the GPC did not help
them commercialise and develop their own businesses.
241. In Assin South enhanced processing of high quality gari production was mainly the result of the
operations of the GPC Dannes Antointed Ltd, which is the only processing centre and monopolizes
the market in the district. Most of the people in the sampled communities in this district are farmers,
not processors. There was no local market for gari in Assin South and those involved in processing
were mostly employees at the GPC. Only those employed at the GPC were exposed to the new
technologies and practices of the GPC and experienced substantial improvements in their livelihoods;
others relied on older processing equipment. The GPC also had its own farm of 350 acres of cassava
to ensure a regular supply of cassava at a moderate price for its factory, which vastly limited the
amount of cassava supplied by the farmers in the supply chain area. The GPC had a very limited
reach with regards to the number of processors that gained access to its improved equipment and
technologies, and the farmers that supplied the center. Some farmers supplied the GPC but were
merely price takers and complained about the price being too low. According to the figures from the
livelihood analysis, in the past 5 years, the number of women involved in processing dropped from
17 to 13% and men dropped from 10 to 8%.
242. In Adansi South, processors gained access to and adopted standardised processing technology and
good management practices, though they did not grow into profitable businesses. Employees at the
GPC were not able to open their own businesses or commercialise; they were strictly employees of
the GPC. The GPC workers said that processing gari at the GPC was a full time job and so they did
not have the time or the financial resources to adopt the new technologies and establish a business on
their own. The GPC was successful to some extent in reaching and teaching local processors and
enhancing the gari processing for the local consumption market. But its trainings declined (from 6 in
2011 to only 1 per year in 2014), and although processors learned a great deal and mostly took up the
quality management practices, smaller and more resource-poor processors have had greater difficulty
in adopting the new technologies.
243. In Nkwanta South, enhanced processing was due to both excess production spilling over into
processing and to GPC production, yet the GPC and other processing centres were not able to expand
and reach more processors due to a failing water and power supply, while trainings and exposure
visits at the GPC have not led to a wider adoption of the improved equipment and technology by
77
processors in the area. The GPC Unity Gari in Nkwanta South was a social-profit enterprise, formally
a cooperative, with a wide interest but limited capacity and reach compared to the districts that had
similar GPCs functioning as social-profits but performing better (e.g. Techiman, Tano North and
Gomoa East). Members of the cooperative were able to increase their processing volumes (form 1 to
3-4 bags per week) due to the trainings and services they received and the use of the improved
equipment at the GPC. They were taught how to hygienically wash and peel cassava, how to store it
for how long, and how to keep good records, and were allowed to use the GPC facilities, which
improved their general business practices and enhanced the quantity and quality of their produce.
This triggered the development of more processing centres (private profits) in the surrounding
communities. Unreliable power supply and inadequate access to water though have seriously
affected the gari processing businesses in this district. Cassava rots because there is no water or
electricity to process into gari. Also trainings and exposure visits that took place at the GPC in the
past five years did not lead to a wider adoption of the improved equipment and technology by smallscale processors in the area. This is because they lacked adequate funds to purchase equipment.
Processors in Odumasi and Krumase (about 5 km each from Krontang) still use the traditional
method of processing cassava. Processors have not been able to package and label their products for
marketing purposes due to the high cost and legal requirements.
244. In the 4 cases (Birim Central, West Gonja, Abura Asebu and Agona East) where enhanced
processing was found weak but fairly attributable to RTIMP (score 3), this was due to the limited
reach or capacity of the GPC. In all of these cases, the adoption and/or use of improved processing
technologies and standardized equipment by resource-poor processors remained very limited.
245. In Birim Central73, the GPC Biakoye operated as a social-profit enterprise where members could
use the facility and equipment for a fee, but its capacity and reach was too small to serve all farmers
and processors. For those who did participate, businesses developed between 2009 and 2012 because
of an increasing demand for cassava and gari, but from 2013 onward, the demand decreased and
prices dropped due to excess production causing market saturation.
246. In West Gonja, processors in one community cluster did use and benefit from improved processing
technologies and equipment to process gari; while in the other cluster (about 19km away) processors
generally still used the manual and traditional methods despite the presence of the GPC Kanyitiwale.
The GPC served as a processing centre for farmers and processors in the district. Mainly female
processors worked to produce high quality, well packaged innovative gari (gari mixed with soya, gari
mixed with margarine). The GPC also served as a location for training and demonstration visits for
processors from Central Gonja and Wa East. Interestingly, this GPC did not reach the processors in
the cluster where it was located. Although some processors had received basic training from OIC and
a grating machine from TechnoServe, the evidence showed that technology transfer and adoption
was not widespread in the area.
247. In Agona East (also gari), none of the evaluated program mechanisms were there74, but cassava
production increased due to the WAAP that provided free planting materials and weedicides, and the
presence of a starch factory and a GPC that initially provided a ready market for farmers’ produce.
Due to local politics75 and financial defaults, both the GPC and the factory had to shut down (see also
73
See also Paragraphs § 178.
According to the MoFA officials and the farmers and processors who attended the district sensemaking workshop, Agona East is
merely an RTIMP-participating district, not an implementing district, which should explain the absence of any RTIMP mechanism.
75
The starch factory was brought in by the NPP (New Patriotic Party) when they were in power. The GPC was established by the
NDC (National Democratic Congress). The minister of agriculture at that time was a very high-ranking member of the ruling NDC.
The GPC was set up in his hometown, Mankrong, and the owner of the GPC is his close relative.
74
78
Paragraph § 149). Meanwhile more than 80 % of the men and women in the Mankrong cluster had
turned to cassava farming and processing as their main livelihood.
248. In Abura Asebu individuals were primarily processing gari whereas the commodity chain was
HQCF. The role of this GPC Tropical Starch was to process HQCF as well as serve as a
demonstration centre for processors in the district. Its reach of processors in the surrounding
communities was scored quite high by the GPC constituent feedback respondents (up to 90% of the
female processors, 30% of male processors, and 80% of the processors younger than 35). Yet the
link (M3b)+C3cC3bO3 was found weak due to the mismatch between the activities of the DSF,
GPC and the processors (see also Paragraph § 137). The GPC organised between 3 and 4
demonstrations a year. Processors were trained on environmental cleanliness, and personal hygiene.
The GPC demonstrations were not on starch and HQCF because these were not relevant for the
people in the community who produced gari. REP trained the Bosomin women’s group on how to
process gari using ginger, soyabean and margarine; CSIR taught them on how to process cassava
flour and soyabean flour and how to use it to make bread, chips, cake, etc. With regards to cassava
flour, the GPC did not have the capacity to fulfil all its orders and contracts and had to turn some
customers away. The GPC owner complained about the lack of regular power which affected
companies in Tema that buy cassava flour which in turn reduced the market for the GPC. Also its
sales of packaged fufu was hindered due to the lack of certification by the FDA.
249. In the 5 districts (Tano North/Dua Yaw Nkwanta, Ho Municipal, North Dayi/Kpando, Wa East and
Ahafo Ano South) where the link (M3b)+C3cC3bO3 was scored very weak with little or no
influence from RTIMP (score 2), there was no access to improved processing equipment and
technologies.
250. In North Dayi / Kpando, there were few processing centres and a dysfunctional GPC. Processing
capacities were quite low. Production of gari and cassava dough was in the decline. Processors were
not motivated due to the lack of market. RTIMP provided equipment for a GPC in 2012, but it was
never used as the installation was never completed. The GPC was a private enterprise with
employees and didn’t organise trainings or demonstrations. Most farmers and processors have never
been to the GPC. The community members did not get along with the owner of the GPC while it was
in operation. No one received business financing or training or market linking services. No farmer
received any financial assistance from a bank and processors did not used improved processing
technology.
251. In Wa East, improved processing technology was not accessible due to the absence of a GPC. Hence
cassava processors were mostly into processing of cassava flour for local food consumption, referred
to as kokonte. This is mainly done manually. The sale of dried peeled cassava is also common in the
area and is mostly done by farmers themselves. Processors were taken on a learning tour to Damongo
to learn how to process gari and cassava flour but the lack of processing equipment in the area
limited processing volumes (see also Paragraph § 191).
252. In Ahafo Ano South, only the GPC Hari Farms processed volumes of high quality gari. No
processor was able to process high quality at scale because they could not afford the equipment and
did not have access business financing to do so. Hari farms GPC is a private enterprise that is
motivated by private profit. The upgrading of this enterprise into a GPC in 2008 did not translated
into processors in the district adopting standardized processing technology and good management
practices. The GPC was not open to farmers and processors to use or learn about its equipment.
79
There is strong evidence that none of the R&T processors developed into independent profitable
processing businesses because farmers do not have the resources to invest in equipment and start a
processing business. Only employees at the GPC were in principle allowed to use its equipment for
processing and this was only when there were no orders running, which was hardly ever the case.
253. In Tano North/Dua Yaw Nkwanta and Ho Municipal, the supply chain leaders were private HQCF
enterprises but not GPCs. In Tano North the factory was not functional however in Ho, the factory
operated whenever it received orders for HQCF (cf. also Paragraph § 151).
254. Generally, where improved processing technologies and standardized equipment were used,
processing volumes and quality significantly increased. Access to these technologies and equipment
was created by introducing cassava processing equipment package and some selected locally
improved technologies to processors, training local artisans to manufacture improved agroprocessing equipment and provide the associated repair and maintenance service, and by upgrading
processing enterprises to GPCs that would serve as demonstration, learning and practice centres
and as market-hubs for processors and farmers. Although the new technologies and equipment have
proven cost-efficient and attractive in terms of its potential return on investments, their adoption
has remained limited due to the overall lack of reach and effectiveness of GPC’s as learning and
good practice centres, and the limited investment capital of small processing centres and individual
processors in the face of saturated markets, failing infrastructure and high inflation.
255. Findings from the cost-benefit analysis of processing equipment conducted in 2014 showed that
following the introduction of improved technologies, access to standardized equipment increased.
The study stated that collaborating manufacturers have reported increasing sales of agro-processing
machines, which shows that more processors have access to technologies needed to pursue their
businesses. The collaborating manufacturers also attributed an increase in the sales of the agroprocessing equipment to the after-sales services they provide, a lesson learnt from RTIMP training.
(MoFA, 2014 a) The increase in sales of improved equipment, however, does not provide any
indication of the scale and magnitude of increased access and adoption, which according to the
evidence obtained from the present impact evaluation as well as the 2014 cost-benefit study remained
very limited.
256. GPCs as well as processors generally agreed that the use of RTIMP equipment saved time, improved
the quality of the products, made the operations and maintenance easier, healthier and more
sustainable, and encouraged increased production. RTIMP promoted the use of standardized
equipment (made from stainless steel materials) by introducing a cassava processing equipment
package and some selected locally improved technologies to processors. This package contains a
peeling bay, washing basin, stainless steel self-feeding grating machine, fermentation bay, presses
(single & double), kiln/stove plus chimney, rectangular stainless steel roasting pans, mechanical
roasting machine, sieve (dry/wet), weighing scales, sewing machine, sealing machine and polytank.
Local artisans were trained in manufacturing this equipment and provide the associated repair and
maintenance service. Furthermore RTIMP upgraded some selected cassava processing facilities into
Good Practices Centers (GPCs), which would serve as demonstration, learning and good practice
centres and as market-hubs for processors and farmers. (MoFA, 2014 a)
257. Also the economic pay-offs in terms of return on investment of the technology have been found quite
attractive. The cost-benefit study have calculated that with the improved gari processing equipment
GPCs can averagely process 15 Mt of fresh cassava per week at a total variable cost of 129,240 GHC
and gross revenue of 235,200 GHC per year, thus leaving them with an average annual gross profit
margin of 105,960 GHC. Non-GPCs and individual processors can process averagely 5.2 Mt of fresh
80
cassava per week at an annual total variable cost of 48,744 GHC and gross revenue of 64,800 GHC,
leaving them with an average annual gross profit margin of GHc 16,056. Thus for every Ghana cedi
revenue generated, GPCs retains 0.45 GHC while non-GPCs retain 0.25 GHC to be put towards
paying fixed cost, showing that gari processing using the improved RTIMP technology and
equipment is a potentially profitable venture for both GPCs and individual processors. This however
only works when there is sufficient market, infrastructure and operational capacity, and fuel and
other costs are not rising too high. (MoFA, 2014 a)
258. However, while there has been a slight increase in the adoption of improved processing technology
and equipment, the cost benefit study of 2014 also acknowledges that this was not widespread.
Aggregated analysis of evidence collected by this PIALA evaluation from the 18 sampled districts
with processing activities showed that in 15 of these, adoption and/or use of improved processing
technologies and standardized equipment by resource-poor processors has remained limited. In all
these cases, the reach and effectiveness of the GPCs as learning and good practice centres was
limited. As is shown in Section 7.2 many also lacked the capacity to innovate, create new demand
and seek new markets.
259. The cost-benefit study conducted in 2014 moreover suggested a lack of investment capital
influenced by gender patterns as an important reason for the limited adoption. (MoFA, 2014 a) As
gender plays an important role in food production and processing in Ghana, post-harvesting and crop
processing is mostly the task of women and children. Most women involved in processing were
resource-poor and had limited capital to invest in processing equipment, which partly explains its
limited adoption of the new equipment has remained limited. This confirmed the observations of the
PIALA researchers.
260. One of the two assumptions of the program’s contribution claim with regard to enhanced processing
was that well-trained processors and farmers would be able to obtain business finance to invest in
their businesses and purchase improved equipment through matching grant funding (MEF). From
both the statistical and the configurational analysis of all the evidence collected in the 25 districts
(including the household surveys), it came out clearly that the majority of resource-poor farmers and
processors received no financial support to invest in their R&T businesses between 2009/2010
and 2014/15, and that by large the MEF was not accessible to farmers and processors, hence did not
make a difference to their adoption of improved technologies and equipment and the development of
their businesses. The evidence for this is presented in the next Section 10.2.
261. Important conditions for the limited investment capital of resource-poor processors (in particular
women) and the limited capacity of GPCs to expand its reach and markets has to do with the failing
markets and infrastructure combined with the rising costs in past years across the entire country (cf.
Section 7.2.2). Where livelihoods were the strongest, this was undeniably the result of a strong
RTIMP presence with strong performing GPCs and better functioning DSFs, which proved that
under more conducive conditions it is possible for resource-poor farmers and processors to adopt
or use the improved technologies and equipment and develop more profitable smallholder businesses,
and for GPCs to become innovative and profitable social enterprises with a wide reach of the
surrounding communities. It is clear that for this to be feasible and sustainable, much more work
need to be done on the market-linking and market creation, infrastructure development, the
development of appropriate regulations with regards to fair competition and trade, and the
development of a mechanism such as the DSF or VC committees that empower all actors involved
in the VC to address issues of unfair competition, breaches of agreement, transportation and supply
& demand linking.
81
262. Finally, a particular constraining condition for GPCs to innovate and expand and thus develop into a
popular market hub at a more advanced stage are the current regulations and procedures for licencing
with the Food and Drugs Authority (FDA) and the Ghana Standards Authority (GSA) which came up
in a few cases (such as Tano North and Abura Asebu). The current regulations and procedures don’t
seem to make it easy for small processing centres to innovate and grow a profitable business, as it
appears too expensive and arduous, and therefore mostly not affordable.
263. Nearly one third of the farmers and processors that were reached by GPCs found that these had
helped them to expand their businesses. Moreover, over half stated that they were able to apply what
they had learned at the GPC. Overall more women were positive than men, and fewer women were
neutral than men. GPC leaders were generally less optimistic in their scoring: only 9 % was positive
about the influence of the GPC on smallholder business development and people’s ability to apply
what they had learned at the GPC.
264. Given the finding mentioned earlier (in 15 of the 18 researched processing districts, GPCs as learning
and good practice centres performed rather weak in terms of reach and effectiveness) it is interesting
to find that 27 % of the people reached by the GPCs believed that the training, support and services
offered at the GPC helped them to expand their processing businesses (score 4-5). Only 10 % was
negative about this (score 1-2), while 63 % remained neutral (score 3). The respondents76 were
farmers and processors that worked or produced at/for GPCs and/or used their facilities to process
their cassava. Women were more positive and less neutral than men: 30 % women were positive
compared to 18 % men; 59 % women remained neutral compared to 74 % men (see figure 10.1.1).
Fig 10.1.1.
265. The GPC leaders were less positive and believed that only 9 % of the farmers and processors had
benefitted from the training, support and services offered at the GPC to the extent that they were able
to develop their own businesses (see figure 10.1.2).
Fig 10.1.2.
76
Respondent distribution table is attached in Annex 4.
82
266. The scores were predominantly from respondents in the gari supply chains (15 in total), as there were
only 2 functional HQCF supply chains in the sample. The scores for gari were almost identical to the
national average scores, with almost no difference between the three zones. Processors in the North
and the South were a bit less negative, and in the North were also consistently a bit more positive,
while in the Centre in particular women were more negative. Young adults (<35 yrs) were generally
less positive in all three zones (5 % in the North, 8 % in the Centre and 16 % in the South).
267. The scores for HQCF were totally different. Since the high quality cassava flour cannot be produced
by individual processors because of the highly sophisticated technology and machinery required,
nobody in the 2 HQCF supply chains was really positive about whether the GPC had helped develop
their businesses. In addition, the GPC leaders scored neutral, thus confirming that the GPC did not
have an effect (positive or negative) on the development of smallholder processing businesses in the
area.
268. The next graph shows that an average of 53 % of the GPC constituent feedback respondents, or the
farmers and processors reached by GPCs, were able to apply what they learned at the GPC (score 45), while only 6 % were not able to apply what they learned (score 1-2), and 41 % were neutral
(score 3). Again women were slightly more positive (4 %) and less neutral (13 %) than men, while
also 8 % of the women scored negative compared to 0 of the men (see figure 10.1.3). Again GPC
leaders were less positive and believed that no more than 9 % of the farmers and processors had been
able to apply what they learned.
Fig 10.1.3.
269. Most people reached by GPCs were women; average 35 % was younger than 35 years. Although
women are in the majority at GPCs, they appear to feel less confident to speak out, express heir
needs and ask for help at the GPCs. Also people younger than 25 appeared to feel less confident.
270. With regard to the GPCs’ reach of women and young adults (<35 years), the charts below stemming
from the KIIs with GPC leaders show that processors patronizing GPCs are predominantly women,
while also in half of the cases processors were younger than 35 years, meaning GPCs have a
relatively good reach of young adults (see figure 10.1.4). Average the reach of young adults comes at
35 %.
83
Fig 10.1.4: Gender and age distribution of GPC participants
271. The graph in figure 10.1.5 below shows clearly that generally more women didn’t feel confident to
express their needs and ask for help at the GPCs than men, despite the fact they are in the majority.
This signals either that women are generally more introvert and solve their problems independently,
and/or that women were more sensitive to power in a male-dominated hierarchy and thus were not
empowered to speak freely at GPCs that are usually managed by men.
Fig 10.1.5.
272. The figures from the GPC constituent feedback analysis furthermore showed that more people
younger than 25 years felt less confident, indicating a sense of insecurity among youth. Obviously
more research is needed in order to draw conclusions about the underlying problems. The evidence
here only points to gender and age-related issues that may require greater attention and thus also
inquiry in the MoFA’s next IFAD-funded program the GASIP.
84
10.2 Changes and causes of access to business finance and market linking
services (link M3c+C1a+M3bC3c)
10.2.1 Changes
273. Resource-poor farmers and processors generally have not been able to access any business finance.
The MEF proved not available and not accessible to the majority of farmers and processors as well
as most GPCs, hence did not make a noticeable difference to their adoption of improved technologies
and equipment and the development of their businesses. Only in a few cases was there evidence of
GPCs and groups of processors and farmers that obtained finance through the MEF or other
channels to invest in their businesses. Farmers and processors attributed negative livelihood changes
mainly to the lack of access to finance.
274. MEF was formally available only in 9 of the 18 researched districts with supply chains that involved
processing activities. In nearly all of these cases, it has not been accessible to farmers and processors.
Also GPCs generally have not been successful in obtaining finance through the MEF. This came out
very clearly from the configuration analysis of all the evidence collected from FGDs and KIIs. Apart
from 2 GPCs in Assin South and Abura Asebu and a few processors in Techiman and Mangpong,
nobody obtained funding through the MEF. In Adansi South and Birim Central, farmers and
processors were able to secure funding for business investments from other sources instead of the
MEF. No other instances were found where processors or farmers, individually or in group, and
GPCs or other small enterprises had been able to access the MEF. The research teams have come
across a few instances of farmers and processors as well as small GPCs who had applied for the MEF
as they fitted the criteria and were encouraged by the MoFA, but none of them were able to secure
the funding.
275. Analysis of the household survey data also showed that only 15 % of the households received some
sort of financial support in the past 5 years to invest in new or existing R&T livelihood activities.
Despite this, 50 % invested in R&T production and 11 % in processing, implying that households
themselves without any financial support or service-provision made most of these investments
(cf. Paragraph § 114-117). Figure 10.1.6 shows that proportionally better-off households received
financial support (18%) than the poorest ones (12%).
Fig 10.1.6: Distribution of households who received finance to invest in R&T businesses in past 5 years
by poverty status
83.3
88.4
100.0
82.1
11.6
50.0
17.9
16.7
0.0
Poorest
No Financial Support
Less Poor
Better off
Received financial support
276. Table 10.1.1 shows that most of the financial support received by households to invest in R&T-based
livelihood activities came from relatives and friends (55%) followed by Susu or local moneylenders
(22%), and much lower from RCBs (14 %). Among households who received support from relatives
and friends, 47 % was less up to 500 while 43 % was in the range of 500-5000 GHS. The finance was
mainly used to increase an existing R&T-based livelihood activity (77%). The same pattern was
observed among all the sources of financial support. For instance, 75 % of households who received
support from relatives and friends used it to increase an existing R&T-based livelihood activity.
85
Table 10.1.1: % distribution of sources of financial support in past 5 years by range of amount
Source of financial support
1 - 500
501 - 5000
>5000
Total
N
Total
Relative and friends
47.2
43.1
9.7
100
72
55.4
Susu or local money lender
35.7
57.1
7.1
100
28
21.5
Rural bank
11.1
72.2
16.7
100
18
13.8
Gov. programs
0
80
20
100
5
3.8
Other sources
57.1
28.6
14.3
100
7
5.4
National
38.5
50.8
10.8
100
130
100.0
277. Average control over the funds was equally distributed among husbands and wives (40 % each).
Table 10.1.2 though shows that funds received from government programs were controlled by the
husbands in 80 % of the households (compared to the wives in only 20 %) and also loans from RCBs
were controlled by the husbands in 50 % of the cases (compared to the wives in only 33 %). On the
other hand, among households who received support from Susu or money lenders, the wives
controlled these resources in 56 % of the cases (compared to the husbands only in 33 %).
Table 10.1.2: Distribution of control of financial support received for investing in R&T businesses
Type of Financial Support
Relative and friends
Susu or local money lender
Rural bank
Gov. programs
Other sources
National
Husband
40.3
33.3
50.0
80.0
16.7
40.6
Wife(s)
36.1
55.6
33.3
20.0
50.0
39.8
Husband and
wife(s)
12.5
7.4
11.1
0.0
33.3
11.7
Other
members
11.1
3.7
5.6
0.0
0.0
7.8
Total
100.0
100.0
100.0
100.0
100.0
100.0
N
72
27
18
5
6
128
278. The SenseMaker analysis of livelihood change narratives (cf. Figure X) showed that in 45 % of
farmers’ and processors’ experiences (n=372), negative R&T livelihood change in the past 5 years
was predominantly attributed to the lack of access to finance, while positive change on the contrary
was rather attributed to access to technology (29 %).
Fig 10.1.7: The change in your story was influenced by …
access to information
access to information
15%
14%
(37)
(19)
19%
15%
(46)
(20)
16%
29%
45%
8%
(38)
(70)
(60)
(10)
access to
finances
access to
technology
access to
finances
access to
technology
86
10.2.2 Causes
279. The MEF was found largely inaccessible because: (a) the mechanism was formally unavailable in
more than half of the cases;(b) the procedure77 for obtaining and paying off MEF funding was too
onerous and often made smallholders pre-invest and sustain operations without sufficient capital or
immediate returns; (c) PFIs showed reluctant to approve applications because of the perceived risk
of investing in farming and agri-processing businesses; and (d) the present conjuncture made R&T
smallholder business investments too precarious for financing.
280. The thematic study of the MEF conducted in 2014 showed an overall positive impact of the MEF on
investment returns, employment and growth of MEF beneficiaries’ businesses. This positive impact,
however, was only observed among those who had received the funding, and not among those who
didn’t receive any funding. In the present PIALA evaluation, MEF was found formally available only
in 9 of the 18 sampled districts with supply chains that involved processing activities (50 %). The
limited availability of the MEF was also confirmed by the 2014 MEF study in which the fund was
found operational in only a few districts. Hence processors and groups in most districts did not have
access to the MEF. This was largely due to the limited participation of PFIs in the program. Banks
and micro finance institutions in Ghana are notorious for being reluctant to provide credit to farming
and agri-processing businesses as these are perceived as too risky. (MoFA, 2014c)
281. The perceived risk of funding investments in R&T farming and agri-processing also affected the
approval of MEF applications. Although the 2014 MEF study reported extremely high approval rates
(between 95 % and 100 %), evidence collected by the PIALA team in the 25 sampled districts
showed that only in 2 districts (Techiman and Mampong) some processors were able to obtain MEF
funding, and only in 2 other districts (Assin South and Abura Asebu) GPCs were able to access the
funding. In 2 districts (Tano North and Upper West Akim), GPCs met the criteria and applied but did
not receive the funding. The configuration analysis revealed no noticeable influence of the MEF on
R&T livelihoods improvements and rural poverty impact nation-wide.
282. In Techiman, one of the six groups of processors (average membership of 22) were able to obtain
financing through the MEF. Although access to finance remained very limited, a positive effect was
recorded on the production of gari, which increased in this district with more than 75%. Exposure to
GPCs combined with business training and (for the happy few) some business financing significantly
contributed to the development of small processing businesses in this district. In Mangpong, MEF
support was given to a few processors before 2012, but after that, nobody had been able to obtain any
more business financing and (apart from some business trainings organised by the BAC) any market
linking services. Even though 10 processors (6 women and 4 men) had been appraised and
recommended for the fund, none of them received the fund. The PFI in the district was depicted as
being cautious in providing MEF credit because of the perceived risks related to farming and agriprocessing. In Assin South, MEF finance was provided by Ecobank only to the GPC. Farmers and
processors in this district were not been able to access the fund or any other financial service, and
haven’t benefitted from the opportunities created by MEF funding of the GPC. In Abura Asebu, the
conditions and criteria for accessing the MEF were found too cumbersome and the GPC was the only
recipient of the loan. No individual or group applied for MEF. Also in Upper West Akim, despite
77
The MEF procedure was the following: The DADU undertook a needs assessment on the prospective beneficiaries as a basis for
possible financing. Upon submission of an application (mostly ranging between 728 GHS to 60,000 GHS), the PFI then inquired if
the potential beneficiary met the requirements. In the case of the Ecobank for instance, processors were required to submit firm
orders with pro-forma invoices and contracts from key customers before loan approval. If the potential beneficiary met all criteria,
then the loan was approved and the application was sent to the RTIMP national office that then granted authorization to transfer the
matching grant component to the requested PFI. Finally, a supplier was paid to manufacture and deliver the requested equipment to
the MEF beneficiary after s/he fulfilled his/her 10 % contribution to the investment.
87
the information and assistance provided by the MoFA, farmers and processors were not able to
access the MEF. Although it met all necessary requirements, the GPC applied but did not receive the
fund and had to wait for more than 2 years to getting its deposit of 10,000 GHS back, causing great
difficulty to continue operations. In Tano North, development of viable businesses was restrained by
the bad road conditions and high transportation costs, which prevented farmers and processors in the
hinterlands from accessing trainings and services. R&T supply chain farmers have not been able to
gain access to business financing and market-linking services even though the MEF was active in this
district. Those who did receive BAC trainings were not able to obtain any business finance. Also the
Adwenepa GPC was not able to access the MEF. Although it met all the criteria and applied in 2011
at the Bomaa Rural Bank and subsequently at the Ecobank, it did not receive the fund. In Adansi
South and Birim Central, farmers and processors were able to access other funding for investing in
their businesses and purchase equipment, but not the MEF. In Adansi South, none of the GPC
employees knew about the MEF and the GPC had not obtained any MEF funding. There was
however knowledge of the EDIF (Export Development and Investment Fund) project that had
operated in the area in 2009 and 2010 and supplied farmers and processors with gari processing
machines. In Birim Central, farmers and processors had knowledge of the MEF but were not able to
access it. Instead they benefited from another fund from the Ministry of Women and Children
(MOWAC) that gave each 500 GHS to expand their processing businesses.
283. Although it’s unclear whether real risk due to inadequate market opportunities rebuked farmers’ and
processors’ access to rural finance, or weather a lack of access to rural finance combined with a
failing rural infrastructure shaped and self-prophesized a perceived risk, it goes without further
saying that the present conjuncture has not been conducive for the sustainable growth of small R&T
businesses. Smallholder investments stand limited chance to succeed in the face of an economic
downturn, failing infrastructure (power, water, roads and market places), insufficient markets,
inadequate finance opportunities and inadequate policies and regulations to support and protect
smallholder businesses. More findings on this can be found in Paragraphs § 182-185 in Section 8.2.
284. Lastly, smallholders generally have limited capital to expand their businesses. According to the 2014
MEF study, this was insufficiently factored in in the design of the MEF mechanism. To meet the
requirement of a 10 % pre-investment78, many MEF beneficiaries used nearly all the capital they had,
hence ended up with improved equipment but little capital for operations. (MoFA, 2014c) The
hierarchy of the PFIs often delayed the process of approval, which on its turn also delayed the supply
of the equipment to the MEF beneficiaries, leaving them to wait without any return on their preinvestment and with little or no capital left to continue their businesses, while loan deduction started
with immediate effect, causing a serious risk for many to default. (MoFA, 2014c) In Upper West
Akim for instance, the GPC had applied for the MEF and put in 10,000 GHS without receiving any
funds or getting back its money for more than 2 years, which caused a financial deficit that affected
GPC operations and payment of farmers.
78
The MEF arrangement required the beneficiary to contribute 10% of the total cost of the improved equipment, while the PFI
provided 50 % in the form of credit and RTIMP/IFAD added 40 % in the form of a matching grant.
88
11 Conclusions and recommendations
11.1 Key findings related to impact
285. In terms of impact on rural people’s access to food and income (I2I1) in the RTIMP treatment
areas, evidence from the statistical analysis of the household surveys and the pattern analysis of the
livelihood micro-narratives have shown three major trends from 2010 until 2015. First, there has
been an increase in access to food and income among rural households, though R&T livelihood
changes did not predominantly affect access to food but rather access to income. Second, 15 % of the
households have raised their income above two dollars (USD) per day between 2009 and 2015,
which largely can be attributed to improvements in R&T livelihoods. Given the context, this can be
considered as a positive impact. Third, as more households moved into R&T farming and
consequently production volumes increased, investments in R&T farming decreased while also
access to technologies decreased, partially due to a shift from production to processing. Investments
and profits from enhanced R&T production and processing remained limited, however, and R&T
livelihood improvements lingered fragile.
286. In terms of changes in livelihoods and the influence of enhanced market linking, production and
processing on this (O1+O2+O3I2), three major findings have come out from the aggregated
configuration analysis of all the evidence collected in the 25 districts. First, in 52 % of the researched
supply chains, improvements of R&T-based livelihoods between 2009 and 2015 were found
relatively strong though not all attributable to RTIMP. The other 48 % generally performed weak in
this regard. In about 32-33 % of the cases, positive as well as negative livelihood changes were
clearly attributable to RTIMP. In a few districts, where RTIMP mechanisms were dysfunctional or
not in place, livelihood improvements were found very weak or virtually absent. Second, FFFs
clearly made a positive difference in 84 % of the supply chains. R&T production boomed across the
country largely due to the introduction of new seed varieties and farming technologies by the RTIMP
and its predecessor RTIP. The new varieties increased the value and volume of raw and processed
R&T produce and contributed to the increase in household income. This initially caused an influx of
people into R&T farming, which led to a substantial increase of production volumes and triggered a
spill over into processing. Third, markets failed to absorb the increased production volumes, which
turned the tide and caused prices to drop, hence negatively affecting farmers’ and processors’
livelihoods from 2013 onward. Accelerated by the economic downturn and poor infrastructure,
inadequate market linking and weak DSF performance in 84% of the cases hampered the growth of
farmers’ and processors’ profits and investments, which rendered improvements in their livelihoods
fragile. Hence the percentage of households that moved up to higher value (and thus income) levels
remained low –only 10 % made it above USD 1/day value and this turned zero at USD 4/day–, while
the percentages of households that did not invest in R&T production (50 %) or processing (89 %) in
the past 5 years were relatively high, indicating a limited profitability.
11.2 Key findings related to market-linking
287. Regarding the changes in market-linking and the influence of the DSFs on these changes
(C1a+M1O), the following four key findings came out from the aggregated analysis of the
evidence collected in the 25 districts. The evidence was found generally quite strong and in most
cases fairly consistent. First, market linking of supply chains through the DSF was found weak and
ineffective in more than 84 % of the researched districts across the country. In 57 % of these, DSFs
to some extent contributed to strengthening the supply chains, but largely failed to link the supply
89
chains to sufficient markets. In 43 % of the cases, DSF contribution to developing the supply chains
was virtually nil and no efforts were made to link farmers and processors to markets. Second, where
supply chain development and livelihood improvements were found relatively strong, despite weak
market linking, this was due to a stronger performance of other RTIMP mechanisms (in particular
GPCs and FFFs) and the influence of other organisations. Where livelihood improvements were
weaker, generally RTIMP and other organisations had a weaker presence and consequently negative
trends such as high inflation and dropping prices exaggerated by poor infrastructure had a bigger
impact on income levels. Where also the supply chains were weak, resource-poor farmers and
processors were much more vulnerable to unfair competition/trade and power abuse by clan leaders
and middlemen controlling the farm-gate prices and the gate to the local markets. Third, only in 16 %
of the researched supply chains, supply chain development and the attraction of new buyers was
comparatively more effective, which enabled farmers and processors to expand their businesses. This
was largely due to a stronger performance of DSFs and GPCs. Both mechanisms were instrumental
in developing supply chains and linking these to new markets. Yet they have not proven strong
enough to withstand external threads and prevent market saturation, due to the GPC’s limited
capacity to innovate and expand, and to the licensing constraints of the Food & Drugs Authority
(FDA) and the Ghana Standards Authority (GSA).
288. Regarding the commercialisation and supply chain linking resulting from enhanced production,
processing and market-linking (M1c+M1b+O3+O2 +O1C1b), there are four major findings. The
first is that R&T production and processing has changed across the entire country from a merely food
producing subsistence to a commercial income-generating livelihood. Both household surveys and
KIIs have confirmed this trend. Evidence collected from districts not treated by RTIMP (such as in
Agona East, Pru and Wassa Amenfi West) showed the necessity of supply chain development and
market linking for enabling smallholders to commercialise. In the absence of any intervention in this
area, resource-poor farmers remain extremely vulnerable to unfair competition/trade. Second,
commercialisation have remained limited and unsustainable in more than 88 % of the researched
districts across the country largely due to market saturation as a result of weak and ineffective market
linking combined with overproduction. Poor roads and poor market infrastructure further limited
resource-poor farmers’ and processors’ market opportunities and in the absence of appropriate
competition regulations rendered them more vulnerable to unfair competition/trade including
monopolistic behaviour of GPCs. Third, in 12 % of the supply chains, commercialisation was found
relatively stronger but inconsistent and not entirely attributable to RTIMP. In these cases GPCs (as
supply chain leaders) have proven an important mechanism to make it possible for resource-poor
farmers and processors (as suppliers) to develop small profitable businesses and gradually grow and
commercialize. The success of this mechanism is largely due to its capacity to innovate and create
new market value/demand, to reach of farmers and processors in the catchment area, and the trust it
can build between the various supply chain actors and their buyers and service providers to establish
strong supply chains.
11.3 Key findings related to production
289. Regarding enhanced production and productivity due the adoption of improved planting materials
and technologies and farmers’ organisation (C2a+C3b O2), two major key findings came from the
configuration analysis of the evidence obtained from KIIs, FGDs and constituent feedback scorings
in 25 districts (which was found strong and consistent). First, R&T production and productivity
increased substantially in about 76 % of the researched supply chains due to the adoption of
improved planting materials and technologies. Where results were rather weak, this was due to a
limited adoption as well as other influences such as beetle infestation, changing weather patterns,
90
limited markets, land tenure, and limited affordability of inputs. Second, there is no evidence of
FBOs effectively helping farmers bargain better prices, fight unfair competition, obtain business
finance, access markets and commercialize. All evidence points to the need for more market
opportunities to enable farmers to commercialise in the first place, and to the need for mixed agribusiness organisations that are less farmer-centred and more attuned to value creation.
290. As for the changes in resource-poor farmers’ and seed producers’ access to improved planting
materials and technologies due to the FFFs (cf. the link M2a+M2b+M2cC2a), there are three
important findings. First, FFFs have proven an effective mechanism to promote the adoption of new
planting technologies and seed varieties, because of its highly participatory character. Widespread
adoption was mainly due to the unsurpassed efficacy of the planting in rows using appropriate
distances and agrochemical application technologies, and the visible benefits in terms of a substantial
increase in quantity/quality and value (in particular for cassava). Second, although women are
generally more involved in cassava production then men and traditionally do most of the work, FFFs
mostly targeted and reached men, in particular small-scale male farmers between 40 and 60 years old
who own a bit of land (max 2 ha). Generally FFFs insufficiently reached women. Third, most FFF
beneficiaries expressed they were able to apply what they learned at FFFs, which helped them
expand their businesses, but young farmers (< 25 years) and women were less positive than adult
men, and also felt less confident to express their needs and ask for help at FFFs. Since in most tribes,
women don’t talk or participate in FFFs, women-specific FFFs should have been organised.
11.4 Key findings related processing
291. For the first link (C1a+M1O), both the statistical and configuration analysis of the evidence
collected in 18 sampled districts where the supply chains involved processing activities, revealed
three major findings. The evidence was collected from the household survey, KIIs, the livelihood
change analyses and the constituent feedback scorings on GPC performance, which all together came
out fairly strong. First, processed volumes of cassava increased considerably in about 50 % of the
gari and HQCF chains (or 9 out of 18) as a result of more people processing cassava and expanding
their businesses by gaining access to training and facilities at GPCs. In only 3 districts (all gari), this
was found robust and attributable to RTIMP due to stronger performing GPCs in terms of market
creation, reach of farmers and processors, and the development of stronger and more inclusive supply
chains. Second, generally where improved processing technologies and standardized equipment
were effectively used, processing volumes and quality increased significantly. These new
technologies and equipment have proven cost-efficient and attractive in terms of its potential return
on investments. Third, adoption/use of improved processing technologies and standardized
equipment have proven ineffective in 15 of the 18 sampled cassava processing districts due to the
limited reach and effectiveness of GPC’s as learning and good practice centres and the limited
investment capital of small processing centres and individual farmers and processors. Both the
household survey and the FGDs undeniably showed limited profits and investments in R&T-based
livelihoods and limited access to financial support to invest in existing or new livelihood activities.
Farmers and processors attributed negative livelihood changes mainly to the lack of access to finance.
Of those reached by GPCs (mostly women, average 35 % < 35 years), nearly one third found that
they had learned at the GPC, thus showing the relevance of GPCs. Women were generally more
positive and less neutral than men, although they appeared less confident to speak out, express their
needs and ask for help at GPCs.
91
292. For the second link (M3c+C1a+M3bC3c), one major conclusion came out of the aggregated
analysis of the evidence collected in the 18 districts (mainly from the household survey, KIIs and
livelihood analyses): the MEF has not proven available and accessible to the majority of farmers
and processors as well as to most GPCs, hence did not make a noticeable difference to their adoption
of improved technologies and equipment and the development of their businesses. The mechanism
was formally unavailable in more than half of the districts, and only in a few cases was there
evidence of groups of processors and farmers as well as GPCs that obtained finance through the MEF
(or other channels) to invest in their businesses. The procedure for obtaining and paying off MEF
funding appeared too onerous, making smallholders pre-invest and sustain operations without
sufficient capital or immediate returns on their investment. Moreover PFIs showed reluctant to
approve applications because of the perceived risk of investing in farming and agri-processing
businesses. Last, the present conjuncture made R&T smallholder business investments too precarious
for financing.
11.5 Conclusions
293. The findings of this evaluation lead to the conclusion that, R&T-based livelihoods initially did
improve between 2009 and 2013, which was relatively strong in about half of the districts and
affected income levels with 15 % of households rising above the threshold of USD 2/day. This
influenced households’ access to food. However, these positive impacts remained limited and
unsustainable largely due to market insufficiency starting from 2013. This was particularly so in
those districts where supply chains and DSF performance were found weak and inadequate, and also
where GPCs did not take up any role in the supply chain linking of small farmers and processors and
did not contribute to the development of their businesses. In these districts, market insufficiency
combined with an inadequate rural infrastructure and land tenure system negatively affected small
and resource-poor farmers’ and processors’ livelihoods and poverty status from 2013 onward, when
the economic downturn struck the country. Evidence points at a reasonable attribution of positive as
well as negative livelihood changes to RTIMP in 32-33 % of the supply chains and of farmers’ and
processors’ individual experiences.
294. The assumption that livelihoods and poverty status could be improved by commercializing
smallholder R&T production and processing businesses, and by developing competitive and
inclusive supply chains, only held true where very strong and concerted efforts were made by the
program partners to: (a) develop solid links between the supply chain actors; (b) address their
capacity and relational issues; (c) create new market opportunities; and (d) expand the catchment
area by widening and deepening the reach of resource-poor farmers and processors in the
surrounding communities. In particular, where the performance of DSFs and GPCs in this regard
were the strongest (12 %), supply chain development and smallholder commercialization was more
successful, resulting in greater livelihood improvements. Where the performance of these
mechanisms were weak, investments in smallholder businesses remained limited and profits stayed in
the hands of a few, thus undermining the hypothesis of smallholder commercialization as the driving
force for sustainable livelihood improvement and rural poverty reduction. However, also in those few
districts with better DSF and GPC performance, livelihood improvements remained fragile due to
insufficient capacity on the part of the GPCs to innovate and expand, further constrained by FDA and
GSA licensing requirements, export regulations, border taxes, and the failing power supply and
infrastructure.
92
295. The assumption that DSFs would help develop sustainable and inclusive R&T commodity chains
largely did not hold true. In 84 % of the sampled districts, DSFs failed to help link farmers and
processors to markets, and in 43 % of these also failed to help establish sustainable and inclusive
supply chains. Lacking were the resources and capacities at the districts (and the higher support
levels) to make this mechanism work –e.g. to conduct proper market analysis and integrated VC
support market creation for smallholder businesses, deepen and expand the reach and role of the DSF,
and propose legislative and policy changes at higher levels needed to make actions at local levels
more successful.
296. The assumption that more resource-poor R&T farmers and processors (including women and young
adults) would commercialize and become part of the supply chains, if they would be able to increase
their production, access markets and develop viable businesses, only held true in the few cases
where these conditions were fulfilled by stronger GPC and DSF performance. Generally, limited
commercialization and ineffective supply chain linking was largely due to:




the limited reach and capacity of DSFs and GPCs to expand, innovate and develop markets;
unfair competition and monopolistic behaviour by traders, entrepreneurs (including GPCs) and
popular leaders (including of MoFA officials);
lack of market opportunities due to a failing rural infrastructure and inadequate policy and
regulations supportive and protective of smallholder business development (including unfair
competition, licencing and certification, export and border tax, etc.); and
lack of trust and investment capital of resource-poor farmers and processors for the above
reasons.
All these causes together hindered resource-poor farmers’ and processors’ ability to commercialize
and enter new markets, and thus outweighed the initial benefits from enhanced R&T production and
processing.
297. The assumption that FFFs would enable R&T farmers and seed producers commercialise by
organising into FBOs and adopting improved planting materials and technologies has proven
partially true. FFFs undoubtedly made a positive difference in 84 % of the supply chains due to
farmers’ massive adoption of the new varieties and technologies, which increased the value and
volume of raw and processed R&T produce and contributed to the increase in household income.
However, counterfactual evidence showed the necessity of strong supply chains and market links to
enable farmers to commercialise. Also no evidence was found of FBOs effectively helping farmers
bargain better prices, fight unfair competition, obtain business finance, access markets and
commercialize. All evidence points to the need for more market opportunities (and thus better roads
and market places, and policies and regulations more supportive of smallholder business
development) to enable farmers to commercialise in the first place, and to the need for developing
mixed agri-business organisations that are less centred around merely farming and more attuned to
value creation (thus including agri-processing and market-linking activities).
298. The assumption that GPCs would reach and teach resource-poor farmers and processors to use
improved technologies & equipment, access business finance and develop profitable businesses, held
true only in a few cases where GPCs more deliberately took up this role (thus functioning more as
social-private profit) and had a greater capacity. Adoption and use of improved technologies and
standardized equipment by resource-poor processors has proven limited in 15 of the 18 sampled gari
and cassava flour supply chains, due to the limited reach and effectiveness of GPC’s as learning
and good practice centres and the limited investment capital of small processing centres and
individual farmers and processors.
93
299. Finally, the assumption that well-trained processors and farmers would be able to obtain a loan
through the MEF to invest in their businesses by large has proven untrue. Resource poor farmers
and processors were unable to access MEF as the mechanism was formally unavailable in over half
of the sampled districts and mostly inaccessible in the other half due to the risks involved.
What were the major barriers for farmers and processors to commercialize and access
markets?
300. The two most fundamental barriers that are conditional for addressing all other limitations are:
 Lack of market opportunities due to a failing infrastructure (in particular power supply, water,
roads and market places);
 Lack of investment capital (only 15 % of the households obtained some sort of financial support
for investing in their R&T businesses in the past 5 years; 45 % of the negative livelihood
experiences were attributed to the lack of access to finance).
301. Another important limitation is the lack of capacity of farmers and processors to organise into
independent and collective agri-businesses that are able to create market value. A more conducive
policy environment and rural infrastructure, however, are conditional to this.
302. From the findings, it came out clearly that more resources and capacities at district and regional
levels are needed to:
 conduct market analysis and develop plans for integrated VC development;
 attract investments for transportation and infrastructure development;
 promote product diversification/innovation and support market/demand creation among GPCs
and other small enterprises with sufficient outreach in the VC catchment areas;
 organise more regular DSF meetings that are open to all supply chain actors and accessible to
more remote communities for discussing market opportunities and issues of unfair competition;
 undertake appropriate action to address the issues raised at DSF meetings and propose changes
in policy and regulations at higher levels needed to make actions at local levels more successful.
303. In the 3 cases where livelihood improvements were found strongest, GPCs were essential to make it
possible for processors to develop profitable business and gradually commercialize. The success of
this mechanism was largely due to its capacity to innovate and create new market value/demand, its
reach of farmers and processors in the catchment area, and the trust it built between the various
supply chain actors and their buyers and service providers to establish strong supply chains.
304. Of those reached by GPCs (mostly women, average 35 % < 35 years), nearly one third found that
they had learned at the GPC, thus showing the relevance of GPCs to resource-poor farmers and
processors (including women and young adults).
What supports or hinders GPCs to better link farmers and processors to markets, and how
is this influenced by the DSF?
305. Most essential impediments identified by this evaluation include:

Limited operational capital
94








Limited capacity to innovate and expand
Failing power and water supply
Expensive licensing and certification procedures
Rising export and border taxes
Rising transportation costs
Limited reach of farmers and processors
Private profit orientation centred on elite interests
Monopolistic behaviour (e.g. unfair price setting, breach of agreements, etc.)
11.6 Recommendations
306. It is clear that RTIMP has made an substantial contribution to the development of opportunities for
resource-poor farmers and processors to improve their lives and livelihoods by turning R&T (the
most important crops grown by the majority of people in Ghana) from a merely subsistence into a
cash crop. This very important shift is largely attributable to RTIMP. Plenty of evidence has been
provided by this impact evaluation that supports this conclusion.
307. Having acknowledged this step forward, there is also the need now for further reflection on the
factors and conditions that have hampered the sustainability of the positive impact that RTIMP has
generated on the lives and livelihoods of the rural poor in Ghana. How can we make sure we draw
the necessary lessons from this successful program in order to enlarge and deepen and strengthen our
impact in the next program (GASIP)? This is the major question for which this impact evaluation
using PIALA hopes to provide evidence-based food-for-thought.
308. We have learned from this evaluation that where livelihood improvements were weaker, RTIMP
(and other organisations) had a weaker presence and consequently negative trends such as high
inflation and dropping prices exaggerated by poor infrastructure had a bigger negative impact on
income levels. We’ve also learned that the new R&T varieties and technologies have caused a
substantial increase in the production and productivity of farmers and the value of their produce,
which caused an increase in household income. The FFFs have been a very effective mechanism in
this regard. Its success though has resulted in excess production that saturated local markets in almost
all districts across the entire country, which hampered farmers’ profits and investments and their
ability to further commercialise. Evidence from districts not treated by RTIMP has sufficiently
proven the necessity of strong supply chains and market links to enable resource-poor farmers and
processors to commercialise. Without sufficient markets, impacts from enhanced production and
processing are unsustainable.
309. Our first critical reflection and recommendation concerns the highly successful FFF mechanism.
Intuitively, everyone would recommend a scaling-up of this mechanism –with adjustments to ensure
greater gender and generation sensitiveness, e.g. by organizing gender- and youth-specific groups.
This definitely would contribute not only to enhancing the value of R&T production, but also to
women’s and young farmers’ empowerment. Being an important source of creativity and innovation
that have remained largely untapped, women and youth (<25 years) would definitely benefit from
their organization into business-oriented farming and agri-processing groups. The FFF concept might
be a suitable mechanism to explore and unleash this idea. However we must inquire and carefully
monitor the conditions that are essential to make this ‘idea’ successful in a conjuncture of rising
inflation and failing markets. Hence we recommend the piloting and scaling up of the formation of
gender- and youth-specific groups with a very careful monitoring of the conditions required to
avoid harm to their livelihoods and trigger the successful growth of these groups into small collective
agri-businesses. These conditions must help resource-poor farmers and processors, and particularly
95
women and youth, overcome the two main barriers to their commercialization identified by this
evaluation, namely: their lack of market opportunities and investment capital.
310. The biggest challenge is market linking. RTIMP performance was generally weak in this area. Given
the circumstances, no blame can be laid on the often-isolated district officials who work in difficult
circumstances. Clearly, there is the urgent need to rethink the DSF mechanism. Commonly DSFs
were organized around the supply chain leaders, mostly small and medium-sized agri-processing
enterprises that were turned into GPCs. In doing so, its reach was limited to the farmers and
processors in these specific supply chains, making them dependent on the supply chain leaders’
benevolence, thus providing the leaders free rein to monopolize the supply chains and the local
markets. The DSF should become a forum that supports inclusive supply chain linking and
encourages innovation and diversification in value creation. By doing so it can provide room for all
farmers and processors and engage them in multiple short and long value chains. Also traders,
transporters, bulkers and off takers need to take part in DSF meetings. For that to happen though,
DSFs must become more successful in the creation of vibrant market linkages. Sufficient resources
and capacities at the districts (and the higher support levels) are needed to make this mechanism
work –e.g. to conduct proper market analysis and integrated VC development planning, attract
private investment, promote product diversification/innovation, support market creation for
smallholder businesses, reach out for farmers and processors and particularly for women and youth to
engage them in the development of strong value chain linkages, and propose changes in policy and
regulations needed to create market opportunities and protect farmers and processors from unfair
competition.
311. The role of ‘supply chain leader’ was taken up by the GPC mechanism. We’ve learned from this
evaluation that GPCs were crucial to make it possible for processors to develop profitable business
and gradually commercialize. The one-third of processors (mostly women, average 35 % < 35 years)
that expressed their satisfaction with the functioning of the GPCs and the benefits they gained has
clearly shown its relevance to resource-poor farmers and processors (particularly women and young
adults). However the success of this mechanism was very limited as it was unclear what is required
to be an effective ‘leader’ in developing strong and inclusive supply chains. Private enterprises were
upgraded to ‘good practice centers’ where demonstrations could be organized. However resourcepoor farmers and processors don’t do the effort to come to demonstrations and enterprises don’t do
social work and reach out for resource-poor farmers and processors, unless they are paid to do so
and/or there is a mutual business interest. Funding obviously is necessary to start the process, yet
doesn’t generate sustainable outcomes in the absence of a genuine business relationship. The
potential power of such a business relationship was shown in a few cases where GPCs functioned as
open social-private profit centres where resource-poor farmers and processors learned to use
improved technologies and equipment and create added value of their produce. Where GPCs were
profitable and attractive businesses in particular for women and young processors, this was largely
due to its capacity to innovate and demonstrate innovation and thus create new market value/demand,
its reach of farmers and processors in the surrounding communities, and the trust it built between the
various supply chain actors and their buyers and service providers to establish strong and inclusive
supply chains. Hence our recommendation here is to expand the concept of GPCs, properly define its
leadership role, and use appropriate performance and feedback monitoring criteria and tools that help
keeping track of the quality and effectiveness of its business relationship with farmers and processors
in the surrounding communities (in particular those resource-poor). Moreover, similar as for the
FFF/FBO-mechanism, careful monitoring is required of the conditions under which GPCs can
become effective leaders of strong and inclusive supply chains. These conditions must help GPCs
overcome the impediments identified by this evaluation, in particular those related to investment
96
capital and market opportunities –e.g. infrastructure, licensing and certification, taxes and
transportation, and policies and regulations.
312. To help GPCs and FFF/FBOs as small collective agri-businesses build their capital and
investments, there is an urgent need for feasible finance mechanisms. Both the household survey
and the participatory research in this evaluation showed a stunning lack of access to financial support
for investing in existing or new R&T business. Commercialisation and adoption of improved
technologies has remained limited in most of the supply chains, not only due to a limited reach and
capacity of GPC’s and DSFs, but mostly due to the lack of finance and market opportunities. The
MEF mechanism attempted to address the issue of finance, yet has largely failed. Its procedure for
obtaining and paying off the loan made it difficult for beneficiaries to pre-invest and sustain
operations without sufficient capital or immediate returns on their investment. Hence the mechanism
needs to be completely restructured in order for it to be accessible to small farmer and processor
businesses. Repayment periods and requirements need to be feasible and agreed upfront. More
thought need to be put in into ‘risk transfer’ and ‘risk distribution’ mechanisms and criteria for credit
worthiness, and into developing different credit packages targeting different categories of businesses
involved in the VCs. Finally, there should be a more comprehensive consultation and communication
process to make all actors involved in the VCs fully understand the risks, the mechanisms and the
requirements regarding repayment and investment. However, to make any credit mechanism work,
feasible business and market opportunities must exist, which in many places in rural Ghana currently
don’t exist.
313. The last recommendation we wish to make is as follows: in order to give all these mechanisms a
reasonable chance to succeed and sustain, much more work is needed on creating an environment
that is more conducive to the growth of small collective agri-businesses. What are the minimal
conditions that need to be in place to trigger growth of profits and investments for these small
businesses? Essential is a minimal rural infrastructure (roads, market places, power and water
supply), which in many places in Ghana is failing. Crucial are also policies and regulations with
regard to fair competition and the use of cassava in end products. A policy that compels industries
producing flour, starch, beer or bio-fuels in Ghana, for instance, to include a percentage of cassava
flour in their products, would definitely spur the demand for cassava. Second, a policy and authority
that regulates competition to make it fair and inclusive, protect smallholder businesses, and prevent
monopolistic practices would certainly aid the DSF and GPC mechanisms to build stronger business
and market relationships between the value chain actors and stimulate greater inclusion of small
farmers and processors. A 2010 study conducted by ODI of the economic impacts of competition in
Ghana (as compared to Kenya, Zambia, Vietnam and Bangladesh) showed that “more competition
with more players, more dynamic entry and exit, and more intense rivalry for customers (…) tend to
deliver better market outcomes” if competition is inclusive and fair, but that in Ghana “market
dominance and anti-competitive practices are very common” while a competition policy and
authority are absent. (ODI, 2010: vi-vii) Competition policies and authorities have proven very
effective in other countries such as Kenya and Zambia “in raising the profile and understanding of
competition issues, in providing evidence and building awareness of the costs of competition
problems, in helping to build up and arm the consumer movement with the evidence it needs to
demand improved market outcomes, and in monitoring market behavior”. (ODI, 2010: 19)
97
314. Finally, as a way to conclude, farmers and processors across the entire country expressed a great
interest in further investment in R&T. Figure 11.6.1 below shows that more than 64 % of the people
who had positive R&T livelihood change experiences in the past 5 years and 52 % of those with
negative experiences, found it’s worth to further invest in R&T; 22% and 31 % respectively was in
doubt; while 2 % and 12 % respectively would rather invest in other crops. This came out from the
SenseMaker analysis of the micro-narratives collected from 393 individual farmers and processors.
Fig11.6.1: Based on my experiences I think it is…
Worth the effort to
further invest in R&Ts
Worth the effort to invest in Worth the effort to
Worth the effort to invest in
crops other than R&T further invest in R&Ts
crops other than R&T
Experiences about positive changes
Experiences about negative changes
(246 stories)
(147 stories)
98
Bibliography and references
Annex B: QA Discussion Points for 1st QA Meeting. (29 January, 2014). Ghana Agriculture Sector
Investment Programme.
Bamberger, M. (2012). Introduction to Mixed Methods in Impact Evaluation (Guidance Note No. 3).
InterAction, Rockefeller Foundation.
Befani, B. (2012). Models of Causality and Causal Inference (Review repaired as part of the DFID study:
Broadening the range of designs and methods for impact evaluation). UK Department for International
Development (DFID).
Bonbright, D., & Power, J. (2010). Private Sector Metrics Contributions to Social Change: Customer
Satisfaction Meets Agriculture Development. IDS Bulletin, 41(6).
Cameron, Sara & Kyei-Mensah, Glowen, (2014). Situation Analysis of Youth and Young People in Ghana,
UNICEF Ghana, Yet to be Published.
Chen, H. T., Donaldson, S. I., & Mark, M. M. (2011). Validity frameworks for outcome evaluation. New
Directions for Evaluation, 2011(130).
Chambers, R. (2003). Participatory Numbers (General Issue No. 47). International Institute for
Environment and Development (IIED).
Chambers, R. (2008a). Poverty Research Methodologies, Mindsets and Multidimensionality. Institute of
Development Studies (IDS).
Chambers, R. (2015). Inclusive rigour for complexity. Journal of Development Effectiveness.
Copestake, J. (2013). Credible impact evaluation in complex contexts: Confirmatory and exploratory
approaches (Draft 18 Oct 2013). Centre for Development Studies, University of Bath.
Creswell, J. W. (2009). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches.
SAGE.
Creswell, J. W., & Clark, V. L. P. (2010). Designing and Conducting Mixed Methods Research. SAGE.
Deprez, S., Huyghe, C., & Van Gool Maldonado, C. (2012). The use of Sensemaker for measuring,
learning and communicating about smallholder farmer inclusion (Case Report). Vredeseilanden-Coopibo
(VECO).
Funnell, S. C., & Rogers, P. J. (2011). Purposeful Program Theory: Effective Use of Theories of Change
and Logic Models. John Wiley & Sons.
Ghana Statistical Service (GSS). (2010). Ghana Living Standard Survey (GLSS). Accra.
Guijt, I. (2008). Seeking Surprise: Rethinking Monitoring for Collective Learning in Rural Resource
Management. Wageningen University and Research Center (WUR).
Holland, J. (2013). Who Counts? The Power of Participatory Statistics. Practical Action.
IFAD (2014). Root and Tuber Improvement and Marketing Programme (RTIMP) Supervision Mission Aide
Memoire.
IFAD (2014). Ghana Agriculture Sector Investment Programme (GASIP) Design Completion Report.
IFAD (2014). Root and Tuber Improvement and Marketing Programme (RTIMP) Supervision Mission
Report.
99
Report.
Report.
IFAD (2010). Root and Tuber Improvement and Marketing Programme (RTIMP) Mid-Term Review
Mission Main Report.
IFAD (2009). Evaluation manual. Methodology and processes.
IFAD (2005). Results and Impact Management System. Practical Guidance for Impact Surveys
IFAD & FAO (2014). Matching Grants with Loans - Experiences and Lessons Learned from Ghana. Rome.
IFAD & BMGF. (2014). Impact Assessment of the ‘Doing Business with the Rural Poor’ Project in Bến
Tre, Viet Nam (2008- 2013) in Ben Tre, Vietnam. Report of the Pilot Application of a Participatory Impact
Assessment & Learning Approach.
IFAD & BMGF (2013a). Potential scalability of PIALA for impact M&E &L. Piloting options and budget
considerations for Ghana.
IFAD & BMGF (2013b). Improved Learning Initiative for the design of a Participatory Impact Assessment
& Learning Approach (PIALA): Insights and lessons learned from the reflections on the PIALA piloting in
Vietnam.
IFAD & BMGF. (2013c). PIALA Research Strategy. Improved Learning Initiative (Internal Document).
Kyei-Mensah, Glowen, (2012). Significant Change Stories, ODI Sponsored Mwananchi Ghana Project.
Levy, S., & Barahona, C. (2002). How to generate statistics and influence policy using participatory
methods in research (Working Paper). Statistical Services Centre, University of Reading.
Macpherson, A. (2014). Root and Tuber Improvement and Marketing Programme Implementation Support
Mission Presentation 3-27 November 2014 Mini Wrap-Up. Kumasi.
Mertens, D. M. (2009). Transformative Research and Evaluation. Guilford Press.
Ministry of Food and Agriculture (MoFA). (2014 a). Assessment and cost-benefit analysis of processing
equipment under RTIMP. Final report.
Ministry of Food and Agriculture (MoFA). (2014 b). Impact assessment of FFF on farmer beneficiaries of
RTIMP. Final report.
Ministry of Food and Agriculture (MoFA). (2014 c). Impact assessment of the MEF in enabling private
sector investment in R&T value chains under RTIMP. Final report.
Ministry of Food and Agriculture (MoFA). (2014 d). Impact assessment of the fresh yam for export value
chain under RTIMP. Final report.
Ministry of Food and Agriculture (MoFA). (2014 e). Tracer study on the effectiveness of business
development and marketing training, and how it has changed behaviour of farmers and processors under
RTIMP. Final report.
Ministry of Food and Agriculture (MoFA). (2014 f). RTIMP 2013 Annual Report .
Ministry of Food and Agriculture (MoFA). (2013). RTIMP 2012 Annual Report.
Ministry of Food and Agriculture (MoFA). (2013). Brief Status of RTIMP as at May, 2013. Kumasi.
100
Ministry of Food and Agriculture (MoFA), (2013). RTIMP Draft Report (November, 2013)- Fresh Export
Yam Value Chain Facilitation.
Ministry of Food and Agriculture (MoFA), (2013). RTIMP Second Interim Report ( July, 2013) –
Consultancy Assignment on Nine Gari for Export Industry Supply Chains –Phase II.
Ministry of Food and Agriculture (MoFA), (2012). RTIMP Final Report (July, 2012) – Plywood Cassava
Flour (PCF) Supply Chain Facilitation.
Ministry of Food and Agriculture (MoFA), (2012). RTIMP Final Report (October, 2012) – Consultancy
Services for Value Chain Facilitation for High Quality Cassava Flour.
Ministry of Food and Agriculture (MoFA). (2012). Brief on Status of RTIMP as at June, 2012. Kumasi.
Ministry of Food and Agriculture (MoFA). (2011). Brief on Status of RTIMP as at June, 2011. Kumasi.
Overseas Development Institute (ODI). (2010). Assessing the Economic Impact of Competition: Findings
from Ghana. London.
Patton, M. Q. (2012). A utilization-focused approach to contribution analysis. Evaluation, 18(3).
Pawson, R. (2013). The science of evaluation: a realist manifesto. London ; Thousand Oaks, Calif: SAGE.
Rihoux, B., & Ragin, C. C. (2009). Configurational Comparative Methods: Qualitative Comparative
Analysis (QCA) and Related Techniques. SAGE.
Rogers, P. (2009). Matching impact evaluation design to the nature of the intervention and the purpose of
the evaluation. The Journal of Development Effectiveness, 1(3).
Stern, E., Stame, N., Mayne, J., Forss, K., Davies, R., & Befani, B. (2012). Broadening the range of
designs and methods for impact evaluations. (Working Paper No. 38). Department for International
Development (DFID).
Van Hemelrijck, A. (2014). Understanding ‘Rigor’: Challenges in Impact Evaluation of Transformational
Development. PhD Research Outline Paper (revised version). Institute for Development Studies (IDS).
Van Hemelrijck, A. (2013). Powerful Beyond Measure? Measuring complex systemic change in
collaborative settings. In Sustainable Participation and Culture in Communication: Theory and Praxis.
Intellect Ltd.
Van Hemelrijck, A. & G. Kyei-Mensah (2014). Design Paper for the impact evaluation of the Root &
Tuber Improvement & Marketing Program (RTIMP). Participatory Impact Assessment & Learning
Approach (PIALA) developed with support of IFAD and BMGF.
Westhorp, G. (2012). Using complexity-consistent theory for evaluating complex systems. Evaluation, 18(4).
White, H. (2009). Theory-based impact evaluation: principles and practice (Working Paper No. 3).
International Initiative for Impact Evaluation (3ie).
White, H. (2014). Current Challenges in Impact Evaluation. European Journal of Development Research
(26:1).
White. S.C. (2015). Qualitative perspectives on the impact evaluation of girls’ empowerment in Banglades.
Journal of Development Effectiveness (7:2).
101
World Food Program (WFP). (2013). Ghana Comprehensive Food security & Vulnerability Analysis
(CFSVA). Accra.
World Food Program (WFP). (2010). Ghana Comprehensive Food security & Vulnerability Analysis
(CFSVA). Accra.
102
Annex 1: Correlation matrix
Annex 1 table 1 on the next page shows the pairwise Pearson Correlation between ‘access to food’, ‘annual
household income’, ‘R&T production/processing’ as main source of income, ‘total value of R&T
production/processing’, ‘total investment in R&T production’, ‘total investment in R&T processing’, ‘total
investment in R&T transportation and other livelihood activities’, ‘access to technology’ and ‘access to
finance’.
The table shows that ‘access to food’ is positively correlated to ‘R&T production/processing’. As more
households moved into R&T production and processing as their main source of livelihood, more
households gained access to food (p 0.27). The relationship is mostly significant (sig.000) though not
direct/linear (p<0.5). Although the increase in ‘total value of R&T production/processing’ has positively
affected access to food (p 0.22), it directly influenced total income increase more directly (p 0.54). Both
correlations are mostly significant (sig 0.000), though a more direct relationship exists between total R&T
value and income (p>0.5). ‘R&T production/processing’ as main source of income is also positively
correlated to ‘annual household income’ (p 0,24), though its relationship with ‘total value of R&T
production/processing’ is more direct (p 0.48), while both are highly significant (sig 0.000). From this, we
can conclude with high level of confidence (>95%) that, as more people turned to R&T as their main
livelihood in RTIMP areas, more people have gained more value out of R&T production/processing in the
past five years, and as a consequence more people have increased their income resulting in more people
having access to food.
‘Total investment in R&T production’ as well as ‘access to technology’, however, are negatively correlated
to ‘production/processing of R&T’ as main source of livelihood. In other words, as more households
depended on R&T, investments in R&T production decreased (p -0.31) while also access to technology
decreased (p -0.24). Both relationships are mostly significant thus strong (sig.000) yet not linear or direct
(p<0.5), meaning other influences were at play. As it appears there is no correlation (p<0.1 ; sig>0.3)
between ‘‘total investment in R&T processing’ and ‘production/processing of R&T’ as main source of
livelihood, this indicates that households with R&T-based livelihoods must have shifted from production to
processing as production increased, clearly indicating a market deficiency.
103
Annex 1 table 1: Correlation Matrix
Production/
Processing of
R&T
Access to
Food
Access to Food
Production/Processing
of R&T
Annual HH Income
Total Value of R&T
Production/Processing
Total Investment in the
R&T Production
Total Investment in the
R&T Processing
Access to technology
Access to Finance
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Total Value
of R&T
Production/
Processing
Annual HH
Income
Total
Investment in
R&T
Production
Total
Investment in
R&T
Processing
Access to
technology
Access to
Finance
1
651
**
.271
.000
651
**
.319
.000
651
**
.221
.000
651
-.061
.132
607
-.027
.506
607
-.004
.924
651
-.005
.907
651
1
837
**
.239
.000
837
**
.479
.000
837
**
-.311
.000
770
.031
.394
770
**
-.237
.000
837
*
.083
.016
837
1
837
**
.543
.000
837
-.049
.172
770
.005
.892
770
-.019
.583
837
.018
.604
837
1
837
**
-.107
.003
770
-.016
.652
770
**
-.141
.000
837
-.007
.836
837
1
770
.068
.061
770
**
.142
.000
770
-.026
.478
770
1
770
**
.220
.000
770
.059
.103
770
1
837
*
.073
.034
837
1
837
104
Annex 2: Inventory of available raw data and sub-reports









Desk review and design workshop report (1)
Aggregated data collation table (1)
National sensemaking workshop report (1)
District sensemaking workshop reports (23)
District data collation tables (25)
Sub-report on household survey analysis
Sub-report on Sensemaker analysis
Constituent feedback analysis spread sheets
Sub-report on synthesis & reflections report
105
Annex 3: Overview of stakeholders interviewed
National and Zonal Stakeholders Interviewed / Consulted
Name
IFAD
Location/District
Accra
Person Interviewed
Andrew Macpherson
IFAD
Accra
Ulac Demirag
IFAD
Accra
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
Kumasi
Kumasi
Kumasi
Kumasi
Kumasi
Freshmacs
Accra
Theophilus Otchere
Larbi
Akwasi Adjei Adjekum
Julius Opuni Asamoah
John Amakye
David Yankey
Angela Osei-Sarfo
Oppong Mensah
Aborampah
Joseph Yeboah
Vincent Akoto
Lambert M. B.
Dandeebo
George Osei-Asibey
Samuel Kwarteng
Nyamekye
Charles Kumi-Amoah
Ecobank Head Office
Ridge Accra
Hanetta Hayford
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
Position
Consultant –
Supervision mission
leader
Country Program
Manager
Country Program
Officer
RTIMP Coordinator
M&E Manager
Yam Supply Chain
Facilitator
Credit Analyst
GPCs and other enterprises
Company Name
Northern Belt
Kanyitiwale
Unity Gari
Aworowa Cassava
Processing Centre
Hansua Women
Gari Processing
Centre
Asueyi
Maxwell Gari
Middle Belt
Hari Farms
Josma AgroIndustries Ltd
Adwenepa GPC
District (clusters)
Person Interviewed
Position
Type &
Functioning
West Gonja (Damango)
Nkwanta South
(Krotang )
Techiman (Aworowa)
Salamatu Braimah
Samuel Amese
Head
GPC - functional
GPC - functional
John K. Amoah
Head
GPC - functional
Techiman (Hansua)
Rebecca Kwartey
Maa
Head
GPC - functional
Techiman (Asueyi)
Kofi Sakyi Daniel
& Faustina Sakyi
None
Head
GPC - functional
None
Enterprise –NOT
functional
Kwaku Sarfo
Mensah
Gibson Senya
Amoah
Mensah
Aborampah
Head
GPC - functional
Manager/
Accountant
Secretary
GPC - functional
West Gonja (Damango)
Ahafo Ano South
(Abesewa)
Ashanti Mampong
(Woraso)
Tano North (Apesika)
GPC - functional
106
Mark & B-face
Bredi Agric
Enterprise
Southern Belt
Biakoye
Charity & Co.
Processing Centre
Jenefal Industry
Ltd
Tropical Starch
Dannes Annointed
Gomoa Obuasi
Agrico Cassava
Processing Centre
Progressive
Women’s
Movement
Bentumah Cassava
Processing Group
Marbet
Adansi South
(Okyerekrom)
Tano North (Dua Yaw
Nkwanta)
Boniface Nickson
Head
GPC - functional
None
None
Enterprise –NOT
functional
Birim Central (Otaipro)
Suhum (Amanase)
Mr. Otoo
Stephen Sekle
Head
Manager
GPC - functional
GPC - functional
West Akim (Adeiso)
Stephen Okyere
Head
GPC - functional
Asebu Kwamankese
(Abura Dunkwa)
Assin South (Assin
Dominase)
Gomoa East (Gomoa
Obuasi)
Alhaji Musah
Head
GPC - functional
Daniel Ankomah
Mends
John Awutey
Agbenorxevi
Head
GPC - functional
Head
GPC - functional
North Dayi/Kpando
(Wusuta)
None
None
GPC – NOT
functional
Agona East (Mankrong)
Aunty Kate
Head
Ho (Akrofu)
Gilbert Asiamah
Head
GPC – NOT
functional
Enterprise (not
GPC)
Zonal research leaders and FFF facilitators
Name
Northern Region
Dr. Stephen Kwaku
Asante
Ashanti Region
F.M Tetteh
Dr. Joseph SarkodieAddo
Dr. Joe ManuAduening
Central Region
Dr. Jonathan Padi
Tetteh
Position
Principal Research Scientist, Savanna Agricultural Research Institute (SARI),
Northern Region
Research Leader, Centre for Scientific and Industrial Research (CSIR) –SRI,
Kumasi
Research Leader, Crop & Soil Sciences Department, KNUST, Kumasi
Research Leader, Centre for Scientific and Industrial Research (CSIR) –CRI,
Fumesua, Kumasi
Associate Professor & Research Leader, Department of Crop Science,
University of Cape Coast (UCC)
PFIs Interviewed
Name
Nkoranza-Kwabre Rural
Bank
Bawjiase Area Rural
Bank
Kwamaman Rural Bank
Ecobank
Location/District
Techiman North
Person Interviewed
Victoria Bonya
Position
Finance Officer
Bawjiase
Wilberforce Kwaku
Adjartey
James Kwarteng
Project/Finance
Officer
Finance Officer
Steven Amoako
Relationship
Manager
Kwamaman, Ashanti
Mampong
Tamale
107
Off-takers Interviewed
Name
Bibiani Logging and Lumber
Company Ltd (BLLC)
Bondplex
Samartex Timber and Plywood
Company
Alhaji Imoro Ibhrahim &Nat
Dasana
Hajia Azara Badawee
Victor Djan
Location/ District
Kumasi Metropolitan
Assembly
Kumasi Metropolitan
Assembly
Wassa Amenfi West
Person interviewed
Mr. Nyameaye
Designation
Manager
Valentine and S. Mensah
Nanumba North
Alhaji Imoro Ibrahim &
Nat Dasana
Hajia Azara Badawee
Victor Djan
Procurement
Officers
Yam Off-takers
East Gonja
Techiman
Yam Offtakers
Cassava off-taker
108
Annex 4: Distribution of participatory research participants
Summary distribution of participatory research participants
Type of method
KIIs with district officials
KIIs with service providers79
Generic Change Analysis with RTIMP intended beneficiaries
Livelihood Analysis with RTIMP intended beneficiaries
Constituent Feedback with DSF participants
Constituent Feedback with FFF participants
Constituent Feedback with GPC participants
District Sense-Making Workshops with RTIMP intended beneficiaries,
participants, district officials and service providers
F
%
M
%
222
189
42
58
79
51
47
49
41
69
217
211
43
84
35
49
53
51
59
31
N
36
39
439
400
85
142
114
640
Distribution of research participants in KIIs
District
Central Gonja
East Gonja
Nanumba North
Nkwanta South
Wa East
West Gonja
Techiman
Adansi South
Ahafo Ano South
Kintampo South
Kumasi Metropolitan
Assembly
Ashanti Mampong
Pru
Tano North
Tano South
Abura Asebu Kwamankese
Agona East
Assin South
Birim Central
Gomoa East
Ho Municipal
Kpando/North Dayi
Suhum
Wassa Amenfi West
West Akim
Community Cluster
Bupei
Sisipe
Bimbilla
Krotang
Gulemga
Damongo 1
Damongo 2
Aworowa
Hansua
Asueyi
Akrofuom
Abesewa
Akora
Kumasi 2
Kumasi 2
Kyeremfaso
Woraso
Zabrama
Dua Yaw Nkwanta
Apesika
Abura Dunkwa
Mankrong
Assin Dominase
Otaipro
Gomoa Obuasi
Ho
Wusuta
Amanase
Samreboi
Adeiso
TOTAL
Zone
NZ
NZ
NZ
NZ
NZ
NZ
NZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
SZ
SZ
SZ
CZ
SZ
SZ
CZ
CZ
SZ
CZ
Officials
3
1
1
3
2
1
0
1
0
0
1
1
2
0
0
2
0
1
1
1
2
2
1
2
2
2
1
1
0
2
36
Service providers
3
2
1
2
0
2
0
4
0
0
2
3
1
1
1
3
0
1
0
2
1
0
1
2
1
2
0
2
1
1
39
79
Service providers also include private actors involved in the program, such as GPC leaders or managers, and local branches of
PFIs.
109
Distribution of participatory research participants in Generic Change and Livelihood Analysis
District
Central Gonja
East Gonja
Nanumba North
Nkwanta South
Wa East
West Gonja
Techiman
Adansi South
Ahafo Ano South
Kintampo South
Kumasi
Metropolitan
Assembly
Ashanti
Mampong
Pru
Tano North
Tano South
Abura Asebu
Kwamankese
Agona East
Assin South
Birim Central
Gomoa East
Ho Municipal
Kpando/North
Dayi
Suhum
Wassa Amenfi
West
West Akim
F
Generic Change Analysis
%
M
%
N
Yapei
Sisipe
Bimbila
Krotang
Gulemga
Damongo 1
Damongo 2
Aworowa
Hansua
Asueyi
Akrofuom
Abesewa
Akora
Kumasi 1
Kumasi 2
NZ
NZ
NZ
NZ
NZ
NZ
NZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
8
7
9
8
9
8
9
7
6
8
7
9
9
0
47
47
47
53
50
57
56
50
43
50
54
53
50
-
9
8
10
7
9
6
7
7
8
8
6
8
9
0
53
53
53
47
50
43
44
50
57
50
46
47
50
-
CZ
0
-
0
-
Kyeremfaso
Woraso
Zabrama
Dua Yaw
Nkwanta
Apesika
Abura Dunkwa
CZ
CZ
CZ
7
8
8
41
57
47
10
6
9
59
43
53
17
14
17
CZ
10
50
10
50
CZ
9
64
5
SZ
10
56
Mankrong
Assin Dominase
Otaipro
Gomoa Obuasi
Giviefe-Ho
Wusuta
SZ
CZ
SZ
SZ
CZ
8
7
8
8
10
CZ
Amanase
Samreboi
Community
Adeiso
TOTAL
Zone
Livelihood Analysis
M
%
%
F
%
100
100
100
100
100
100
100
100
100
100
100
100
100
8
9
9
11
8
5
9
7
8
8
10
7
8
0
50
47
47
52
50
38
60
47
57
47
50
64
50
-
8
10
10
10
8
8
6
8
6
9
10
4
8
0
50
53
53
48
50
62
40
53
43
53
50
36
50
-
N
%
16
19
19
21
16
13
15
15
14
17
20
11
16
100
100
100
100
100
100
100
100
100
100
100
100
100
0
-
0
-
100
100
100
2
7
0
22
54
-
7
6
11
78
46
100
9
13
11
100
100
100
20
100
0
-
8
100
8
100
36
14
100
2
33
4
67
6
100
8
44
18
100
5
38
8
62
13
100
57
47
57
50
53
6
8
6
8
9
43
53
43
50
47
14
15
14
16
19
100
100
100
100
100
8
9
9
8
5
53
60
64
50
45
7
6
5
8
6
47
40
36
50
55
15
15
14
16
11
100
100
100
100
100
5
42
7
58
12
100
8
53
7
47
15
100
CZ
10
56
8
44
18
100
8
47
9
53
17
100
SZ
10
53
9
47
19
100
7
47
8
53
15
100
CZ
0
222
51
6
217
100
49
6
439
100
100
4
189
40
47
6
211
60
53
10
400
100
100
17
15
19
15
18
14
16
14
14
16
13
17
18
110
Distribution of participatory research participants in DSF, FFF and GPC Constituent Feedback (CF) FGDs
District
Central Gonja
East Gonja
Nanumba North
Nkwanta South
Wa East
West Gonja
Techiman
Adansi South
Ahafo Ano
South
Kintampo South
Kumasi
Metropolitan
Assembly
Ashanti
Mampong
Pru
Tano North
Tano South
Abura Asebu
Kwamankese
Agona East
Assin South
Birim Central
Gomoa East
Ho Municipal
Kpando/North
Dayi
Suhum
Wassa Amenfi
West
West Akim
Community
cluster
Zone
CF with DSF participants
F
%
M
%
N
%
CF with FFF participants
F
%
M
%
N
%
F
CF with GPC participants
%
M
%
N
%
Yapei
Sisipe
Bimbila
Krotang
Gulemga
Damongo 1
Damongo 2
Aworowa
Hansua
Asueyi
Akrofuom
Abesewa
NZ
NZ
NZ
NZ
NZ
NZ
NZ
CZ
CZ
CZ
CZ
CZ
3
5
5
0
4
0
0
0
0
0
0
0
43
56
56
50
-
4
4
4
0
4
0
0
0
0
0
0
0
57
44
44
50
-
7
9
9
8
-
100
100
100
100
-
3
4
3
4
5
0
4
2
4
4
0
1
38
50
33
50
56
44
25
50
44
14
5
4
6
4
4
6
5
6
4
5
0
6
62
50
67
50
44
100
56
75
50
56
86
8
8
9
8
9
6
9
8
8
9
7
100
100
100
100
100
100
100
100
100
100
100
0
0
0
7
0
6
0
7
5
7
8
0
78
67
100
62
88
100
-
0
0
0
2
0
3
0
0
3
1
0
0
22
33
0
38
12
0
-
9
9
7
8
8
8
-
100
100
100
100
100
100
-
Akora
Kumasi 1
Kumasi 2
CZ
CZ
CZ
0
0
0
-
3
0
0
100
-
3
-
100
-
2
0
0
40
-
3
0
0
60
-
5
-
100
-
0
0
0
-
0
0
0
-
-
-
Kyeremfaso
Woraso
Zabrama
Dua Yaw
Nkwanta
Apesika
Abura Dunkwa
CZ
CZ
CZ
CZ
4
0
0
0
57
-
3
0
0
0
43
-
7
-
100
-
4
0
0
0
57
-
3
0
0
0
43
-
7
-
100
-
5
0
0
0
36
-
9
0
0
0
64
-
14
-
100
-
CZ
SZ
0
4
50
0
4
50
8
100
4
3
57
38
3
5
43
62
7
-
100
-
2
3
50
38
2
5
50
62
4
8
100
100
Mankrong
Assin Dominase
Otaipro
Gomoa Obuasi
Giviefe- Ho
Wusuta
SZ
SZ
CZ
SZ
SZ
CZ
0
0
7
2
3
0
88
25
38
-
0
0
1
6
5
0
12
75
62
-
8
8
8
-
100
100
100
-
0
1
4
0
0
0
17
57
-
0
5
3
0
0
0
83
43
-
6
7
-
100
100
-
0
5
7
5
0
0
62
88
62
-
0
3
1
3
0
0
38
12
38
-
8
8
8
-
100
100
100
-
Amanase
Samreboi
CZ
SZ
0
0
-
0
0
-
-
-
2
0
40
-
3
0
60
-
5
100
9
0
100
-
0
0
-
9
-
100
-
CZ
5
42
50
49
5
43
50
51
10
85
100
100
4
58
50
41
4
84
50
59
8
142
100
100
3
79
50
69
3
35
50
31
6
114
100
100
Adeiso
TOTAL
111
Distribution of participants in the district sensemaking workshops
Zone
Officials &
Service
Providers
F
%
M
%
N
%
F
%
M
%
N
%
Yapei
Sisipe
Bimbila
Krotang
Gulemga
Damongo 1
Damongo 2
Aworowa
Hansua
Asueyi
Akrofuom
Abesewa
Akora
Kumasi 1
Kumasi 2
NZ
NZ
NZ
NZ
NZ
NZ
NZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
CZ
7
4
3
5
1
5
0
7
0
0
3
6
2
0
0
7
5
3
8
8
10
0
11
0
0
17
6
12
0
0
44
42
23
47
47
48
48
53
55
44
-
9
7
10
9
9
11
0
12
0
0
15
5
15
0
0
56
58
77
53
53
52
52
47
45
56
-
16
12
13
17
17
21
0
23
0
0
32
11
27
-
100
100
100
100
100
100
100
100
100
100
-
0
6
4
0
0
0
0
0
0
0
0
0
0
0
0
86
80
-
5
1
1
1
1
0
0
0
0
0
0
0
0
0
0
100
14
20
100
100
-
5
7
5
1
1
-
100
100
100
100
100
-
Kyeremfaso
Woraso
Zabrama
Dua Yaw
Nkwanta
Apesika
Abura Dunkwa
CZ
CZ
CZ
CZ
4
0
2
2
9
0
9
21
60
50
40
6
0
9
31
40
50
60
15
18
52
100
100
100
0
0
8
0
67
-
0
0
4
0
33
-
12
-
100
-
CZ
SZ
2
3
9
13
38
52
15
12
62
48
24
25
100
100
0
0
-
0
0
-
-
-
Mankrong
Assin
Dominase
Birim Central
Otaipro
Gomoa East
Gomoa Obuasi
Ho Municipal
Giviefe- Ho
Kpando/NorDayi
Wusuta
Suhum
Amanase
Wassa Amenfi
Samreboi
West Akim
Adeiso
TOTAL
SZ
SZ
3
5
15
17
54
65
13
9
46
35
28
26
100
100
0
0
-
0
0
-
-
-
CZ
SZ
SZ
CZ
CZ
SZ
CZ
3
5
3
3
5
0
7
90
11
12
3
12
12
0
17
247
58
50
19
55
46
49
48
8
12
13
10
14
0
18
272
42
50
81
45
54
51
52
19
24
16
22
26
35
519
100
100
100
100
100
100
100
0
0
0
0
0
0
0
18
42
0
0
0
0
0
0
0
13
58
31
100
District
Central Gonja
East Gonja
Nanumba North
Nkwanta South
Wa East
West Gonja
Techiman
Adansi South
AhafoAno South
Kintampo South
Kumasi
Metropolitan
Assembly
Ashanti
Mampong
Pru
Tano North
Tano South
Abura Asebu
Kwamankese
Agona East
Assin South
Community
Cluster
Farmers & Processors
Transporters & Traders
112
Annex 5: Overview of national sensemaking workshop participants
National level
Name
Akwasi Adjei Adjekum
Julius Opuni Asamoah
John Amakye
Angela Osei-Sarfo
Oppong Mensah Aborampah
Joseph Yeboah
Vincent Akoto
George Osei-Asibey
Samuel Kwarteng Nyamekye
Osei Tutu
Obrien Nyarko
Roy Ayariga
Represenative
Ulac Demirag
Theophilus Larbi
Sarah Ashu-Davis
Franck Luabeya Kapiamba
Daniel Pasos
E.K Nkansah
Angela Dannson
Abdul Abdulai-Rahman
Hanetta Hayford
Prof Jonathan Padi Tetteh
Dr Joseph Sarkodie-Addo
Wilberforce Kwaku Adjartey
Victoria Bonya
Prosper Yaw Klu
Yakubu Iddris Sheref
Fordjour K. Minka
Lambert Abusah
Emmanuel Garti
Dr. Emmanuel Moses
Dr. Stephen K. Asante
Dr. F. M. Tetteh
Robert Arthur
Samuel Forson
Benedict Tuffuor
James Kwarteng
Theophilus Larbi Tetteh
Organisation
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
RTIMP
MOFA, Mampong
MOFA, Sekyere Central
GASIP
GASIP Climate Change
IFAD
IFAD
IFAD
IFAD
IFAD
MOFA
MOFA, Accra
Ecobank, HQ Accra
Ecobank , HQ Accra
UCC
KNUST
Bawjiase Area Rural Bank
Nkranza-Kwabre Rural Bank
Brim Central Desk Officer
NRGP
RDA, Ashanti Region
PPMED, MOFA
PPMED, MOFA
CRI, Kumasi
CSIR-SARI, Nyamkpala
CSIR-SRI
Asuansi Agric Station
LD Consult, Accra
TREND, Accra
Kwamanman Rural Bank
Akatachiman Rural Bank
Northern zone
Name
Loverage Amenu
Atchulo Azara
Julius Gamegah
Langa M. Thomas
District
Nkwanta South
Central Gonja
West Gonja
East Gonja
Designation
Desk Officer
AEA
Desk Officer
Desk Officer
113
Augustine Akono
B. A. Samsun
Margaret Kumah
Faustina Ama Ameyaa
Samuel Amese
Mrs. Faustina Sakyi
Tsetsefle Doris
Dokurugu Tandana
John K. Amoah
Salamatu Braimah
Techiman
Central Gonja
East Gonja
Techiman, Hansua
Nkwanta South
Techiman, Asueyi
Nkwanta South
Nanumba North
Techiman
West Gonja, Damongo
Danso Christiana
Nkwanta South
Nanumba North
West Gonja, Sori 3
Alhaji Imoro Ibrahim
Jaliatu Moro
Desk officer
Processor
Farmer/trader
farmer/processor
Processor
processor/farmer
Processor
Yam Farmer/ Service Provider
Farmer/Service provider
Processor
Farmer
Yam Farmer/trad
Processor
Southern zone
Name
Cecilia Gboloo
W.H. Fordjour
Felix Agbenyega
Asante Andrews
Samuel Naah
John Agbenorxevi
Stephen Okyere
Alhaji
Anastasia Fynn
Hawa Otwe
Samuel Okyere
Juliana Aboagyewaa
Nicholas Dwamena
Alfred Agbordzor
Gilbert Asiamah
Issah Killanu
Christopher Kofi
District
Designation
Kpando
Abura-Asebu-Kwamankese
Ho Municipal
Upper West Akim
Assin South
Gomoa East
Upper West Akim
Gomoa East
Suhum
Suhum
Ho Municipal
Ho Municipal
Ho Municipal
West Akim
DDA
DDA
Desk Officer
Desk Officer
Dannes Annointed Ltd GPC
Agrico Cassava Processing Center GPC
Jenefal Industries Ltd
Tropical Starch
Processor
Farmer/ Processor
Farmer
Farmer/ Processor
Farmer
Processor
Processor
Farmer
Name
District
Designation
Yaw Okyere
Thomas Fofie
Ali Umaru
Asher Owusu Toku
Mr. Opoku
Fatima Salifu
Grace Addai
Lydia Nyantakyiwaa
Dwomoh Boateng
Osei Yaw Benett
Eno Mary
Adobea Ama
Kofi Waja
Mampong
Mampong
Ahafo Ano South
Adansi South (New Edubiase)
Ahafo Ano South
Ahafo Ano South
Mampong
Mampong
Mampong
Mampong
Mampong
Kintampo South
Kintampo South
Desk Officer
BAC Officer
AEA
Desk Officer
Cassava Farmer
Cassava Farmer
Cassava Farmer
Cassava Farmer
Gari Processor
Cassava Farmer
Gari Processor
Yam Farmer
Yam Farmer
Central zone
114
Niyangnan Wajabu
Michael Ndema
Paul Amoah
Suya Sona
Josma Agro-Industries (Janet)
Afia Jiga
Emmanuel D. Kuwornu
Asomani Odei-Mensah
Pru
Kintampo South
Pru
Mampong
Pru District
Kumasi
Yam Farmer
Yam Farmer
Yam Farmer
Yam Farmer
Yam Trader
RTIMP
DDA
Research & facilitation team
Name
Glowen Kyei-Mensah
Nana Sikaba King
Kobby Optson
Essi Haffar
Darlene Okai
Ama Gyan
Abubakar Mohammed
Bernard Alando
Samuel Entee
Anthony Amuzu
Johnathan Anaglo
Organisation
PDA
IFAD
PDA
PDA
PDA
PDA
PDA
PDA
PDA
SNV
GSS
University of Ghana
115
Annex 6: Sampling frame
The sampling frame used is the frame of the list of the commodity supply chains from the four main types of the
commodities in the RTIMP market system stratified into agro ecological zones. Table 1 shows that there are 67
catchment districts of the commodity supply chains of the 4 types of commodity chains developed by RTIMP.
Table 2 shows the relative distribution of the total number of the supply chain leaders per type of commodity
chain. Table 3 presents the number of districts sampled (25 in total) that has been proportionally allocated to
each zone. Six districts were systematically selected in the North zone, 12 in the Central zone and 7 in the South
zone. Table 4 presents the numbers of distribution of the proportional allocation of commodity supply chain
areas among the sampled districts. Table 5, finally, presents the numbers of distribution of the proportional
allocation of commodity supply chain areas among the 3 main agro-ecological zones covered by the program.
The samples are stratified and selected independently in two stages from this sampling frame. In the first stage,
25 districts will be selected with probability proportional to size (PPS) selection procedure according to the
sample allocation given in Table 3. Implicit stratification with proportional allocation will be achieved at each of
the lower administrative unit levels by sorting the frame before the sample selection according to a certain
geographical order, within each of the explicit stratum, and by using a probability proportional to size selection
procedure.
After the selection of districts a list of the distribution of the various types of the commodity supply chain in
each district has been prepared. This list helps identify the communities and locations of the commodity supply
chains in each of the selected districts.
A fixed number of 30 locations/clusters have been proportionally allocated to each commodity using a
probability proportional to size (PPS) according to the sample allocation given in Table 4. These locations form
the centres of the supply chain areas in which qualitative interviews and household surveys will be conducted.
In each of the supply chain areas, 30 households will be systematically selected among three communities that
form the cluster, i.e. approximately 10 in each community with equal probability.
Table 1: Distribution of the number of catchment districts of the commodity supply chains
Zones
Name of zone
1 North (Savannah)
2 Central (Transitional & deciduous)
3 South (Deciduous)
Number of catchment districts of the commodities
14
36
17
67
Total
Table 2: Distribution of the number of supply chains for each type of commodity chain
1
2
3
4
Type of commodity
PCF Chain
High Quality Cassava Flour (HQCF)
Fresh Yam Export chain (FYE)
Gari Chain
Supply chain population
Total
8
7
7
29
51
Table 3: Distribution of the proportional allocation of districts among the zones
116
Zones
1
2
3
Total
Name of zone
North (Savannah)
Central (Transitional & deciduous)
South (Deciduous)
Allocation of districts to agro-ecological zones
6
12
7
25
Table 4: Distribution of the proportional allocation of commodity supply chain areas among the sampled districts
Zones
1
2
3
4
Total
Type of Commodity
PCF Chain
High Quality Cassava Flour (HQCF)
Fresh Yam Export chain (FYE)
Gari Chain
Number of supply chains
5
4
4
17
30
Table 5: Distribution of the proportional allocation of commodity supply chain areas among the zones
Zones
1
2
3
Total
Name of zone
North (Savannah)
Central (Transitional & deciduous)
South (Deciduous)
PCF
HQCF
1
3
1
5
0
2
2
4
FRESH YAM GARI
2
2
0
4
4
9
4
17
117
Annex 7: Poverty distribution tables
Poverty distribution of the households
Wealth Status
Poorest
Less Poor
Better Off
Total
Frequency
Percent
335
162
340
837
40,0
19,4
40,6
100,0
Poverty distribution of the households by Zone
ZONE
Wealth Status
Northern
Poorest
Less Poor
Better Off
Central
47
40
97
184
Southern
197
76
151
424
Total
91
46
92
229
335
162
340
837
Poverty distribution of the households by Region
Regions
Poorest
Northern
Upper West
Ashanti
Brong Ahafo
Eastern
Central
Volta
Western
Wealth Status
Less Poor
37
29
6
6
23
6
129
55
45
15
61
26
23
20
11
5
335
162
Better Off
Total
58
18
30
90
31
47
49
17
340
124
30
59
274
91
134
92
33
837
Poverty distribution of the Male households by Region
Regions
Poorest
Northern
Upper West
Ashanti
Brong Ahafo
Eastern
Central
Volta
Western
Total
Wealth Status
Less Poor
36
25
6
4
15
4
90
47
21
13
34
21
21
13
10
4
233
131
Better Off
Total
51
14
25
65
22
36
40
15
268
112
24
44
202
56
91
74
29
632
Poverty distribution of the Female households by Region
Region
Northern
Upper West
Ashanti
Brong Ahafo
Eastern
Central
Volta
Western
Total
Poorest
Wealth Status
Less Poor
1
0
8
39
24
27
2
1
102
Better Off
4
2
2
8
2
5
7
1
31
Total
7
4
5
25
9
11
9
2
72
12
6
15
72
35
43
18
4
205
118
Poverty distribution of the Male households by Zone
Zone
Poorest
Northern
Central
Southern
Total
Wealth Status
Less Poor
45
33
126
64
62
34
233
131
Better Off
Total
81
112
75
268
159
302
171
632
Poverty distribution of the Female households by Zone
Zone
Northern
Central
Southern
Total
Poorest
Wealth Status
Less Poor
2
71
29
102
Better Off
7
12
12
31
Total
16
39
17
72
25
122
58
205
119
Annex 8: Household survey questionnaire
IMPACT SURVEY QUESTIONNAIRE
CONSENT.Hello. My name is
.We are conducting a survey with IFAD and the Ministryof
.This survey will help us investigate changes in people’s lives in areas where cassava, yam and
other roots and tubers are produced and sold. You are invited to participate in this survey voluntary. You can choose not
to answer any questions, and you can stop the interview at any time. All of your responses will be confidential. Would
you like to ask me anything about the survey? Do you agree to participate in this survey?
Respondent agrees
YES
to be interview
NO
Exit
END
Continue
QUESTIONNAIRE IDENTIFICATION
ZONE: _______________________________
CODE ___________________________
REGION NAME: _____________________________________
CODE_________________
DISTRICT NAME: ____________________________________
CODE_________________
COMMUNITY CLUSTER NAME: _______________________
CODE_________________
INTERVIEWER NAME ________________________________
CODE_________________
NAME OF HOUSEHOLD HEAD: __________________________________
HOUSEHOLD NUMBER:_________________
DATE OF INTERVIEW: ___ ___ / ___ ___ / ___ ___ ___ ___
TOTAL NUMBER OF MEMBERS IN THIS HOUSEHOLD: _________________
OUTCOME OF INTERVIEW
Completed
Incomplete
Absent
Refused
Could not locate
1
2
3
4
5
120
SECTION 1: HOUSEHOLD SOCIO-DEMOGRAPHICS
No.
01
02
03
04
05
06
07
08
09
10
11
12
Please tell me the first name of each person who usually lives here, starting with the Head of the Household.
List adult members of the household first, then list children.
AGE
SEX:
HIGHEST LEVEL OF EDUCATION
How old was (name)
Male= 1
FIRST NAME:
on his/her last
Female = 2
For only member who are 6 years or older
birthday?
Never
Primary Secondary
attended
Tertiary Grade Completed at
/JHS
/Diploma
NAME
M
F
AGE
school
that Level
(Head of Household)
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
1
2
1
2
3
4
121
SECTION 2: SURVEY QUESTIONS
NO.
QUESTIONS AND FILTERS
CODINGCATEGORIES
A. ACCESS TO FOOD
A.1
Did your household experience periods of
food shortage during….?
YES
NO
(Yes / No)
A.1.a
A.1.
b
A.2
The past year 2014
1
2
The year before (2013)
1
2
Have any of the children under 15 in this
household experienced illness in the past 12
months?
YES
1
NO
N/A
2
3
(Yes / No)
Once
1
2
3
4
Specify
3
4
bottled water
Other
(Pond, river or stream / Boor holes / Pipe water /
Sachet or bottled water)
2
Sachet or
1
Pond, river or
What is the main source of drinking water
for your household?
Three or more times
Never
(Once/ Twice/ Three or more times)
A.4
Twice
Pipe water
How often have children under 15 in this
household been put to sleep at night without
sufficient food (thus remaining hungry) in the
past 12 months?
stream
Bore hole
A.3
5
Husband and
5
6
7
8
Growth of Yam
1
2
3
5
6
7
8
Growth of Cocoyam
1
2
3
5
6
7
8
Growth of other roots & tubers
(specify)
1
2
3
5
6
7
8
B.1.e
Processing of Cassava into High Quality Cassava
Flour (HQCF)
1
2
3
5
6
7
8
B.1.f
Processing of Cassava for Plywood industry
1
2
3
5
6
7
8
Processing of Cassava for Gari
1
2
3
5
6
7
8
B.1.
b
B.1.c
B.1.
d
B.1.
g
wife/s
Other
Wife/s
3
B.1.a
important
2
Who controls the income?
(Husband / Wife / 2nd or 3rd Wife / Husband and wife/s / Other)
important
3rd most
1
(most important / 2nd most important / 3rd most important)
Most
Growth of Cassava
What are the MAIN sources of income
for your family?
important
2nd most
Husband
B.1
members
Specify
B. ECONOMIC ACTIVITIES, INCOME & PROFIT
122
Processing of Cassava for other use
(specify)
1
2
3
5
6
7
8
B.1.i
Trading and transportation of Fresh Yam for Export
1
2
3
5
6
7
8
B.1.j
Trading and transportation of other Root & Tubers
1
2
3
5
6
7
8
B.1.
k
Other types of livelihood activity
(specify)
1
2
3
5
6
7
8
2
3
4
5
B.2.
b
The year before (2013)?
1
2
3
4
5
B.2.c
Five years ago (2009)?
1
2
3
4
5
B.2.a
contributed
Not R & T
501-1000
1001-2000
2001-5000
5001
In what range did the value of TOTAL R&T
production (including from farming and
processing)
of your household fall in …?
members who
0-500
B.3
household
The past year 2014?
1
2
3
4
5
6
B.3.
b
The year before (2013)?
1
2
3
4
5
6
B.3.c
Five years ago (2009)?
1
2
3
4
5
6
(0-500, 501-1000, 1001-2000, 2001-5000, or 5001 or more GHS
/ year)
B.3.a
New
Existing
Husband
Wife/s
Husband and
wife/s
Other
In the past 5 years, has your household
established a new, or increased an existing,
livelihood activity that forms a RELIABLE
source of income?
If yes, who has done most of the new or extra
work?
No
Skip to C.1
B.4
household
Production of Cassava
1
2
3
5
6
7
8
Production of Yam
1
2
3
5
6
7
8
Production of Cocoyam
1
2
3
5
6
7
8
Production of other roots & tubers
(specify)
1
2
3
5
6
7
8
B.4.e
Processing of Cassava into High Quality Cassava
Flower (HQCF)
1
2
3
5
6
7
8
B.4.f
Processing of Cassava for Plywood industry
1
2
3
5
6
7
8
Processing of Cassava for Garry
1
2
3
5
6
7
8
B.4.a
B.4.
b
B.4.c
B.4.
d
B.4.
g
E
(Husband / Wife(s) / Husband and wife(s) / Other)
Specify
5001
1
(0-500, 501-1000, 1001-2000, 2001-5000, or 5001 or more GHS
/ year)
Number of
and above
2001-5000
The past year 2014?
0-500
1001-2000
In what range did the TOTAL income of your
household fall in …?
How many members of the household
contributed to this total income?
501-1000
B.2
and above
B.1.
h
123
B.4.
h
Processing of Cassava for other use
(specify)
1
2
3
5
6
7
8
B.4.i
Trading and transportation of Fresh Yam for Export
1
2
3
5
6
7
8
B.4.j
Trading and transportation of other Root & Tubers
1
2
3
5
6
7
8
B.4.
k
Other types of livelihood activity
(specify)
1
2
3
5
6
7
8
Wife/s
(specify type)
Husband and
wife/s (specify
type)
Other
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
Specify
Husband
(specify type)
2
(specify type)
5001 and above
(specify type)
1
(Range 0-5000 or more)
Who controls most of these financial resources
and decides when and for what to use them?
501 – 5000
(specify type)
Financial assets:
In the past 5 years how much financial support
has your family received to invest in these new, or
increase these existing, livelihood activities?
1 - 500
(specify type)
C.1
0
C. ASSETS
C.1.a
C.1.
b
C.1.c
C.1.
d
From relatives and friends
Specify whether it is for new or existing livelihood
(2=new, 3=existing)
From Susu or local money lender
(2=new, 3=existing)
C.1.e
From local rural bank
(specify)
C.1.f
(2=new, 3=existing)
C.1.
g
Government programs such as RTIMP
(specify)
C.1.
h
(2=new, 3=existing)
C.1.i
Other sources
(specify)
C.1.j
(2=new, 3=existing)
C.2
Productive assets:
How many of the following items does
your household currently owned?
(Number)
C.2.a
Land plots ( convert to the nearest acres)
C.2.
b
Farm equipment – hoes and cutlasses
C.2.c
Farm equipment –animal drawn plow
C.2.
d
Farm equipment –tractor-drawn plow
C.2.e
Farm equipment –power tiller
124
Farm equipment –other
(specify)
C.2.
h
Processing equipment
(specify)
C.2.i
Storage for agricultural produce
C.2.j
Other productive assets
(specify)
C.3
Household assets:
How many of the following items does
your household currently have?
0
1
2
3
or
>3
0
1
2
3
or
>3
Specify
C.2.
g
(Number)
C.3.e
Cell phone
C.3.f
Bicycle
C.3.
g
Motorcycle or scooter
C.3.
h
Car or truck
C.3.i
Other means of transportation
(specify)
C.3.j
Water storage
C.4
Knowledge assets:
In the past 5 years, in how many of the following
training or demonstrations did someone from
your family participate?
Who decides whether/when to use most of what is
learned or obtained in these
trainings/demonstrations?
Specify
Refrigerator
Other
C.3.
d
Television
Husband and
wife/s
C.3.c
Radio
Wife/s
C.3.
b
Electricity
Husband
C.3.a
C.4.a
C.4.
b
C.4.c
C.4.
d
Training on R&T production technology
Training on R&T seed multiplication
Training on R&T trading and financial
recordkeeping
Demonstration of cassava processing equipment
and quality management
THANK YOU VERY MUCH FOR YOUR CO-OPERATION.
ADDITIONAL NOTES OR COMMENTS: ____________________________________________________________
125
Annex 9: Key Informant Interviews Questionnaires
Participatory Financial Institutions (local Branch Manager assigned)
Overarching evaluation/learning questions:





To what extent and under which conditions have resource-poor processors and farmers (in
particular women and young adults) been able to access financing through the MEF to invest in
their businesses?
Who has the MEF been able to reach, and who not, and what is the most plausible explanation for
this?
What are the effects of the MEF on growth of the funded agro-processing businesses?
What has been done to avoid elite capture?
To what extent have Participating Financial Institutions (PFIs) been prepared to accept MEF
applications and provide credit to resource-poor processors and farmers (in particular women and
young adults) up to 50% of their planned investments? Under what conditions?
KII Questionnaire:
1. How does the MEF work within your institution?
 What are your specific functions in relation to the RTIMP?
2. What is your catchment area? (District, community?)
3. Who qualifies to apply for funding? What are the criteria? (e.g. received business training;
resource-poor…)
4. How do you verify these criteria?
5. How many people have applied for funding?
 How many of these were males, females, younger than 35 years old? Explain why
6. How many have received funding and why?
 How many of these were males, females, younger than 35 years old? Explain why
 What percentage of those who received MEF were resource-poor farmers? How do you
know? (probe)
 What were the barriers to receiving funding?
 What were the barriers to accept MEF applications and provide credit to resource-poor
farmers and processors?
 What was done to overcome these barriers?
7. Do you monitor how funding is used?
 How do you monitor this? What are the monitoring mechanisms you use?
8. To what extent do you think your funding has contributed to the growth of small R&T agroprocessing businesses in the area? Explain.
**Ask for any document that might help to find out the number/overview of all those who have taken loans
126
FFF Facilitators and Research Leaders
(You may not find the research leaders in the districts, since they are regionally based, but if they are there,
they also should be interviewed)


To what extent have FFF enabled resource-poor R&T farmers and seed producers (including
women and young adults) to commercialize and organise themselves?
Have FFFs sufficiently reached the more vulnerable and/or illiterate resource-poor farmers
(including women and young adults)? What have FFFs done to help them overcome barriers to
participate?
KII Questionnaire:
1. What are your specific roles with respect to R&T?
2. How are the FFFs organised?
 Who participates in them?
3. In the past year how many FFFs has your team facilitated?
 In how many communities has your team conducted FFFs?
 How frequent were the trainings your team facilitated?
 Who is targeted? (specify characteristics of target groups)
 Who participated? (specify characteristics of participants)
 Were females and young farmers (under 35 years) targeted in any of your trainings? How
do you target them?
 COMPARE WITH CONSTITUENT FEEDBACK:
On average, what percentage of your participants were women and men, and what
percentage were young farmers (under 35 years) in the past year? Have these percentages
changed in the past 5 years? Explain.
 Compared to the last 5 years, has there been an increase or decrease in training and
participation? Why?
4. In your opinion, do your trainings sufficiently reach more vulnerable and/or illiterate resource-poor
farmers and seed growers (including women and young adults)?
5. After the trainings, do R&T farmers and seed producers have access to improved production
technologies?
6. What kind production technologies have they received? Can you provide an overview?
7. How are these improved production technologies distributed? Who is targeted?
8. How many individuals have received improved production technologies? Can you provide a list?
9. Have R&T farmers and seed producers been able to adopt the improved production technologies
(explain)? How are you able to verify this?
10. Have R&T farmers and seed producers been able to afford the improved production technologies?
How do you know this?
11. What are some of the barriers that prevent the adoption of improved production technologies?
12. COMPARE WITH CONSTITUENT FEEDBACK:
To what extent did the trainings help farmers expand your farming businesses?
 Please give a score from 1 to 5
 Why/why not?
How confident do you thing all farmers are to express their needs and ask for help at an FFF
meeting?
 Why/why not?
127

How is this different for women than for men? What about young farmers?
To what extent did the training help farmers expand your farming businesses?
 Why/why not?
Have farmers been able to produce more at lower cost past/two/three years ago through participating
in the FFF?
 Yes / No
 Why/why not?
** Obtain list of FFFs organised with thematic areas, and review RTIMP productivity surveys and progress
reports from the SRID, GLDB, DDAs and ZOCs.
Off-Takers
Overarching evaluation/learning question:


What are the main barriers to linking resource-poor farmers and processors (including women and
young adults) to old and new R&T commodity markets?
What conditions need to be in place to help them overcome these barriers? What is missing (e.g.
certification, packaging, traceability, market prospection)?
KII Questionnaire:
1. What is the process of acquiring R&T raw materials?
 What are some of the challenges you encounter in acquiring some of these raw materials?
2. Have you had new suppliers of R&T raw materials in the past 5 years?
3. To what extent is the produce of the standard/quality that you need? Explain.
4. To what extent is the produce of the quantity that you need? Explain.
5. Are your R&T raw materials delivered in a timely manner? Why/why not?
6. How did you establish your linkages with the raw material producers?
7. Do you know if women and young adults are part of your raw material producers?
8. In your opinion, what are the main barriers to linking small farmers and processors to your
industry? Have you done anything to help overcome these barriers?
9. Are you aware of R&T programmes? Which ones?
10. Are you a part of any of these programmes?
 If yes, which one(s)?
 Are there any advantages to being a part of these programmes?
11. Have you ever attended a District Stakeholder Forum (DSF)? If yes, were they beneficial to you?
Why/ why not?
District Officials
(including District Head, DDAs, RTIMP Desk Officer, Extension officer)


To what extent have the DSF contributed to the development of sustainable and inclusive
commodity chain linkages?
Who participated in the DSF –who did not? What motivated farmers and processors to participate
(or not)?
128

What enables or thwarts DSFs to become viable “chambers of commerce” –i.e. member networks
that serve as private business linking and market-information platform empowering buyers,
producers and processors (including women and young adults) to address their demand & supply
issues independently?
KII Questionnaire:
1. What is your role in RTIMP?
1. Who are the primary target groups of RTIMP?
 How do you define resource-poor?
2. Which institutions are involved in the organisation of the DSFs?
 What are the respective roles of MOTI and MOFA in organising the DSFs?
 Did the persons in charge include MOTI, MOFA in the organisation?
 How is the collaboration between these institutions?
3. How are DSFs supposed to help establish Commodity Chain linkages? To what extent, and how,
did DSFs contribute to the establishment and maintenance of market linkages?
 What is the role of DSF? Why did it work well, why didn't it work well?
 What markets existed for resource-poor farmers in this district before 2009? What markets
existed for processors in this district before 2009? Has it changed? How? What additional
markets have been included, if any?
 What linkages existed before the DSFs?
 What are the problems encountered in establishing links with these new markets? Any
advantages?
 What needs to be done to overcome these problems? What is missing? (e.g. certification,
packaging, traceability, market prospection)?
 To what extent has the capacity-building (business development and marketing)of
resource-poor farmers and processors contributed to the development of commodity chain
linkages?
 In your opinion what is preventing strong linkage of resource-poor farmers and processors
to old R&T markets? What about to new R&T commodity markets?
4. How are the DSFs organised?
 What were the issues discussed at each DSF?
 Which actors patronise the DSFs? (specify characteristics of actors)
 Is there a fair representation of different actors at the DSFs? (buyers, producers and
processors) Why/why not? Explain.
 How are participants invited/selected to participate in the DSFs?
 If you had the opportunity to include some things to make a positive change in how the
DSFs are organised, what would they be? Why?
 In your opinion what are the main barriers to the positive change your would like to see?
5. According to you, what makes DSFs function well?
 To what extent do they function as independent member networks?
 COMPARE WITH CONSTITUENT FEEDBACK:
To what extent are decisions made at the DSF carried out in practice by all the actors?
o Please give a score from 1 to 5
o Why/why not?
 To what extent do they help buyers, producers and processors (including women and young
adults) to address their demand & supply issues independently?
 In your opinion, who do you think benefits most from participating in the DSF? And in
what way?
 In your opinion what enable/thwarts the DSF’s to become viable chambers of commerce?
6. Who participates in DSFs?
 How frequent have DSFs been organised in the past 5 years?
 Who participated? (specify characteristics of participants)
129





To what extent are the resource-poor farmers and processors actively engaged in the DSFs?
What are the barriers that exist for the resource poor farmers and processors to actively
participate in the DSF’s? Explain.
In your opinion what are the reasons for low/high attendance of women/young adults?
Have females and young farmers (under 35 years) been encouraged/targeted to participate?
What are the barriers for them to participate? What has been done to overcome the barriers
for them to participate?
COMPARE WITH CONSTITUENT FEEDBACK:
On average, what percentage of the DSF participants were women and men, and what
percentage were young farmers (under 35 years) in the past year? Have these percentages
changed in the past 5 years? Explain.
Compared to the last 5 years, has there been an increase or decrease in participation in the
DSF? Why?
How confident do you think resource poor farmers and processors are or feel to express their views
and ask for support/assistance at the DSF?
 Why/why not?
In your opinion, have farmers and processors been able to buy/sell more in the past one/two/three
years through participating in the DSF?
 Yes / No
 Why/why not?
** Review:
 RTIMP Enterprise Record Books (ERBs);
 ZOCs progress reports;
 MoFA and DADU Organisational Capacity Assessments;
 RTIMP M&E data.
Business Advisory Centres (BAC’s)

To what extent has the training organised by the BACscontributed to the capacity building needed
to establish and sustain Commodity Chain linkages? Probe for barriers, opinions, reach,
enablers**
KII Questionnaire:
1. Approximately, how many resource-poor farmers and processors participated in the trainings last
year?
 Who qualifies to be part of the trainings? (specify)
 How many were women? How many were men? How many younger than 35?
 How has this changed over the past 5 years?
 What is done to reach more women and young farmers (<35 years)?
2. Can people easily come to the trainings? Are there difficulties? What is done to help those living
far and have less possibility to attend the trainings?
3. To what extent are your trainees also participating in DSF’s? Why/why not? Explain.
 In your opinion, to what extent does your training (business development and marketing)
contribute to enabling farmers and supply chain processors solve their supply and demand
issues?
4. To what extent has the training organised by the BACs contributed to increased access to business
financing?
 How many of your participants have applied for MEF?
130

Were they successful? How many and why?
To what extent have your trainings helped improve the profitability of the businesses of your
clients?
 Why/why not?
Good Practice Centres (GPC)
(including leader and financial manager)





Who has been engaged in/exposed to the GPC, and who not?
To what extent have GPCs been able to reach and teach resource-poor processors about good
quality processing and management practices, including the use of improved technologies and
standardized equipments?
How and to what extent has participation in GPCs affected the processors’ ability to develop more
profitable agri-processing businesses?
What conditions need to be in place for GPCs to become profitable and attractive businesses
particularly for young adults living in remote areas?
What supports or hinders GPCs to better link the supply chain farmers to old and new markets, and
how is this influenced by the DSF?
KII Questionnaire:
1. When and how did your enterprise become a GPC?
2. How is the GPC managed? What is the organizational structure?
3. Does your GPC receive any funding or subsidies or other financial support? Is this support needed
to survive or can the GPC stand on its own?
4. Do you organise demonstrations?
 If yes, how often annually?
 What is the purpose of the demonstrations?
 Have the numbers of participants increased or decreased in past years? (Can you supply a
list of participants)
 Who attends your demonstrations? (specify characteristics of participants)
 How are people invited/selected?
 What would you have hoped to see as an impact/outcome of your demonstrations on
people’s farms and businesses?
9. In what other ways do farmers and processors engage with the GPC?
 Who participates? (specify characteristics of participants)
 What are the barriers that exist for resource poor farmers and processors to actively engage
with GPC? What has been done to overcome these barriers?
 Have females and young farmers (under 35 years) been encouraged/targeted? What are the
barriers for them to actively engage with the GPC? What has been done to overcome these
barriers?
On average, what percentage of the GPC participants were women and men, and what percentage
were young farmers (under 35 years) in the past year? Have these percentages changed in the past 5
years? Explain.
To what extent do you think resource poor farmers and processors have been able to obtain the
support, training and services they need through your GPC in the past 2-3-4-5 years in order to
expand their processing business?
131


Please give a score from 1 to 5
Why/why not?
To what extend do you think resource-poor farmers and processors feel confident to express their
needs and ask for help from the GPC?
 Why/why not?
To what extent do you think small resource-poor farmers and processors have been able to adopt
and apply what they learned at your GPC?
 Why/why not?
To what extent do you think that the training support and services provided through your GPC have
helped resource-poor farmers and processors to expand their processing business?
 Why/why not?
Have they been able to process more at lower cost in the past three years through participating in the GPC?
 Yes / No?
 Why/why not?
Have they been able to sell more in the past three years through participating in the GPC?
 Yes / No?
 Why/why not?
17. Have you heard of MEF?
 Do the people who participate in your demos know about the MEF and know what it is?
Has anyone applied?
 In your opinion, who has access to MEF funds?
 In your opinion, how have MEF funds affected resource-poor farmers’ and processors’
ability to develop more profitable agri-processing businesses?
18. What conditions need to be in place for your GPC to become more profitable and attractive
particularly for young people (<35 years)?
19. What has hindered or supported your GPC to find new buyers and better link the small farmers and
processors to new markets in past 5 years?
 How has this been influenced by the DSF?
132
Annex 10: Generic Change Analysis
A. What, When and Why to use this method
The Generic Change Analysis is a PRA-inspired method that uses two tools in focus group discussions to
inquire a program’s impact claim –namely: change ranking, followed by a causal flow mapping. The data
collected from this method is qualitative and complementary to the quantitative data from the household
survey.
The change ranking is a descriptive data collection tool that seeks to identify and rank the main changes in
roots- & tubers-based livelihoods of the past 5 years in terms of their impact on people’s wealth &
wellbeing as defined by the beneficiaries themselves. It takes the form of a brainstorm, thus is rapid and
brief. The process essentially involves the identification, specification and ranking of key changes.
Subsequently the causal flow mapping inquires the possible explanations by taking the one or two changes
with greatest impact (thus highest rank) as a starting point to map out their impacts and causes, link these
back to RTIMP, and collect detailed information on who has been affected most/least and why. This
exercise takes more the form of an in-depth discussion involving the participants in the reconstruction and
visualization of the causal flow.
For the evaluation of RTIMP, the focus of this method is on changes in roots- & tubers-based livelihoods
affecting rural poor people’s access to food and income and their ability to work and access services. This
concerns the links between the O’s and the I’s in the ToC diagram.
B. With Whom to use this Method
The generic change ranking and causal flow mapping will be conducted with separate gender-specific
groups with a good representation of young people in the age range between 15 and 34, randomly
sampled from the communities in the community clusters in the sampled districts. Each group will be
composed of 10 people (either women or men), of which 6-8 from the primary target groups (R&T
processors and farmers) and the others from the wider group of rural poor that was targeted by the program.
C. Time and materials
The total time needed to do the whole exercise should take no more than two hours. This includes the time
needed for discussion and bathroom checks.
The materials you need at hand for doing the exercise should include: 3-5 flipcharts, 2-3 colors of
markers,2-3 colors of stick-it notes or moderation cards, tape, card board (the size of a flipchart), and a
camera.
D. Step-by-step guide
Introduction and ethics clearance (10 min)
1. Welcome the group, thank them for their time, explain the purpose of the exercise and ask people’s
agreement to participate.
-
The purpose is: to conduct research on the changes in the past 5 years that had an impact on
rural livelihoods, what caused these changes, and how it affected people’s wealth and
wellbeing in the communities.
133
-
Do BY NO MEANS mention the program being evaluated and do NOT say this is an
evaluation.
-
Explain clearly we’re trying to understand how people themselves experienced and valued
these changes, and NO DIRECT BENEFITS will be given, other than the opportunity to
engage in the discussions around these changes at the local and national levels.
-
If asked, explain that the research is co-funded by multiple partners who are interested in
learning from rural people’s experiences and views through this research.
2. Complete the group profile by letting everybody introduce him/herself and noting down the full
names of the participants (see note-taking template).
-
If there is objection, propose to use nicknames.
3. Ensure safety, anonymity and confidentially.
-
Build trust and make sure participants feel safe and comfortable before starting the exercise.
-
Ask permission to take photos and make sure these photos cannot be connected to any specific
quotations.
-
Ask outsiders (or those not invited –including officials and leaders) to leave the group before
starting the exercise.
4. Be aware of power dynamics in the interactions –particularly between researchers and researched,
leaders and non-leaders, poorer and better-off, older and younger, higher and lower educated, and
other social and cultural differences that may prevent participants from expressing their views.
-
Carefully take notes on all participation issues and power dynamics occurring.
-
Reflect with your team every evening on how to best deal with these issues and dynamics.
-
If there is any risk or threat, interrupt the exercise immediately.
5. Output:
-
Group profile filled in.
Tool 1 – Generic Change Ranking (40 min)
6. Start with a quick brainstorm to identify the main characteristics of ‘wealth’ and ‘wellbeing’ and
different levels of wealth and wellbeing in the communities.
-
Capture the main characteristics for instance of those least, middle, more and most wealthy and
well off.
-
Avoid getting caught up in too much detail –just make sure you create a common
understanding and categorization.
7. Next ask participants to think about what has changed in the past 5 years related to roots and tuber
production and marketing, which positively or negatively affected the wealth and wellbeing of most
people in their communities.
-
Prompt for CHANGES (not problems) in R&T-based livelihoods (not in wealth & wellbeing)
which had an effect on wealth & wellbeing of MOST PEOPLE (not just their own households).
If participants come up with a problem or activity, ask what it has done to their household and
the wider community, to identify the change.
-
Use a timeline if needed to trigger people’s memory. Keep asking: “What other changes can
you think of that happened last year, the year before, in 2010, in 2009, or in 2008?”
-
Write positive and negative changes on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!) of two different colours.
134
-
Continue the brainstorm for 10-15 minutes or until there are at least 2-3 positive and 2-3
negative changes formulated on cards that had a significant effect on people’s wealth and
wellbeing.
-
If there are many, cluster those that are similar or closely related, and give the clusters a name.
8. Next, ask the participants to rank the cards or clusters in terms of magnitude of impact on wealth
and wellbeing of the people in their community.
-
Start with asking “which of these changes had the biggest impact on wealth and wellbeing of
the people in your community”?
-
Ask how big the impact was, for whom the impact on wealth & wellbeing was the biggest
(specific categories), and how many % of the people in their community were affected
seriously.
-
Next ask for the change with the second biggest impact and ask how big the impact was, for
whom biggest and % people affected?
-
Compare each new change with the former one, and rank them based on biggest impact on
most people, and mostly affecting the least and middle wealthy & well-off people.
-
Note that understanding why certain changes had a greater impact than others, and for whom, is
more important than the ranks in themselves!
-
At the end of the ranking, verify the order by picking the two cards at the top of the list and
asking “Just to make sure we got it right: Is it correct that this change has had the biggest
impact on the majority of people and particularly on those who previously were not so wealthy
and well off?”and working your way down through all the cards repeating this question.
9. Finally, conclude the ranking with a brief discussion how much the changes on the cards listed at
the top (those with the biggest impact) have made those who were 3-5-7 years ago less wealthy and
well-off. This forms the basis for the causal flow mapping.
10. Outputs:
-
List of characteristics of wealth & wellbeing for different categories of people (from least to
most wealthy and well-off)
-
List of changes in R&T-based livelihoods that positively or negatively affected people’s wealth
and wellbeing, ranked according to biggest impact on most people in the communities, and
mostly felt by those more vulnerable and less wealthy and well-off.
-
Explanation of why changes affected some groups more than others, in some places more than
others, and how much it changed people’s wealth and wellbeing status.
Tool 2 – Generic Change Causal Flow Mapping (40 min)
11. Take the 2 highest ranked changes (one positive and one negative) and tell the participants that the
discussion now will focus on these two changes.
-
If the selected changes are rather vague, ask the participants to make it more specific. For
example, instead of ‘credit’, you will have a clearer discussion with a change formulated as
‘access to affordable business start-up loans’.
-
Placing these cards in the middle of a flipchart –the positive above the negative– on a table or
the ground (use a hard board underneath the flipchart, if needed).
12. Probe first for the impacts of each change. Use the same colours as the ones used in the former
exercise for indicating whether the impacts are positive or negative.
-
Ask: “what has happened as a result of this change? how has it affected people’s lives?”
135
-
Write positive and negative impacts on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!) of two different colours, and place the cards on the right side of the
changes in the middle.
-
Probe for enough details (what, when, where, for whom, how many, why) to make it possible
to triangulate with other data (for instance from the household survey).
-
If someone mentions an indirect consequence, ask what happened before in order to link it to
the original change.
-
Probe for second and third line impacts by asking: “What has happened as a result of this
change/impact? And what has happened as a result of that?”
-
Continue probing until you have exhausted all impacts and have reached the level of “sufficient
access to food and income to lead and sustain a healthy and active life”
13. Next probe for the causes of each change in the middle of the flipchart. Ask for the causes of these
changes, and the causes of the causes (first, second and if needed also third line causes).
-
Ask questions like: “What has made these changes possible? What else contributed to or
influenced this change? What were the underlying causes of each of these direct causes?”
-
Write positive and negative causes on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!) of two different colours, and place the cards on the left side of the
-
Probe for enough details (what, when, where, for whom, how many) to make it possible to
verify with other sources. For instance, in case of ‘better infrastructure’ probe until you arrive
at something as specific as ‘the new bridge over the canal that links the village to the next
district and was built just before the latest elections’.
-
Keep on probing until exhausted and have reached the level of services and interventions.
14. Finally, ask the participants if all changes, causes and impacts are placed in the right order in
relation to each other and ask them to help you draw the connections.
15. Output:
-
Generic causal change flow diagram (hard copy in the notes).
Closure (10 min)
16. Thank participants for their time.
17. Ask permission to take a photo of the diagram, share it with others and use it for discussion at the
district (together with other groups’ diagrams).
18. Ask who would be interested to participate in the district-level feedback discussions and represent
the group’s findings.
- Make sure those less influential or articulate participants get an opportunity to volunteer!
19. Finally, ask the participants for feedback at the end the exercise, about what they learned and found
most interesting and useful in doing the exercise, and how they would want to use the findings that
came out of this session.
20. Output:
-
Photo of the Generic causal change flow diagram.
-
List of people who want to participate in the district-level sensemaking workshop.
-
Description of the outcome for the participants of the exercise.
136
Annex 11: Livelihood Analysis
A. What, When and Why to use this method
The livelihood analysis is a participatory method that uses three tools in focus group discussions to
investigate the changes in livelihoods caused by the program being evaluated, or by other influences –
namely: change matrix, change signification, and causal flow mapping. The data collected from this
method is mainly qualitative, although systematic data collection and the use of sensemaker techniques also
permit quantitative analysis of patterns of perceptions and experiences across larger populations.
The change matrix is a descriptive data collection tool that helps to obtain an overview of the different
types of livelihood activities in the communities and the major changes that have happened in these
livelihood activities in the past 5 years, as well as women’s and men’s engagement in each of these and the
relative income and risk levels. For this it uses PRA-based ranking, proportional piling and scoring
techniques.
The change signification tool uses techniques borrowed from the patented SenseMaker approach80 to help
surface patterns (both expected and unexpected) and provide an additional layer of quantified qualitative
data collected in a systematic manner across larger populations.
Finally, the causal flow mapping is an explanatory data collection tool that maps out the impacts and causes
of the one or two most significant changes in the R&T livelihood activities, link these back to RTIMP, and
collect detailed information on who has benefited (or not) and why. This last exercise takes the form of an
in-depth discussion involving the participants in the reconstruction and visualization of the causal flow.
For the evaluation of RTIMP, the focus of this method is on changes in roots- & tubers-based livelihoods
caused by the program mechanisms and other influencers, and affecting rural poor people’s access to food
and income and their ability to work and access services. This merely concerns the links between the C’s
and the O’s in the ToC diagram.
B. With Whom to use this method
The livelihood change analysis and causal flow mapping will be conducted with separate gender-specific
groups with a good representation of young people in the age range between 15 and 34, randomly
sampled from the communities in the community clusters in the sampled districts. Each group will be
composed of 10 people (all women or all men) that are direct program beneficiaries (R&T farmers, seed
growers and processors).
C. Time and materials
The total time needed to do the whole exercise should take no longer than two and a half hours.
The materials you need at hand for doing the exercise should include: 3-5 flipcharts, 2-3 colors of markers,
2-3 colors of stick-it notes or moderation cards, tape, card board (the size of a flipchart), a box of stones or
seeds, and a camera
80
Cf. http://www.sensemaker-suite.com.
137
-
-
evaluation.
-
-
-
-
-
quotations.
-
-
-
-
25. Output:
-
Top 5 facilitation tips
1. Organize the livelihood analysis sessions in the mornings when it’s not too hot, and if possible on
days that people have a bit more time. Prepare your flipcharts with triads and the diad (including
the examples) in advance. Also prepare some of the cards with types of livelihood activities in
advance, in particular when you expect most of the participants are illiterate. This will save you lots
of valuable time –not only your time but also and most importantly the time of the participants!
2. KEEP UP THE PACE‼ This is a lengthy method so you need to make sure that you get through the
first tool in NO LONGER than 50 minutes. If you don’t, you will run into trouble –meaning: you
will not be able to finish the causal flow mapping so will have to organize an EXTRA FOCUS
GROUP the next day to do the causal flow mapping (with other participants, and starting from a
new brainstorm on most significant changes in livelihood activities).
138
3. When explaining the tool using the examples of a triad and diad, ask the participants to stand
around the flipcharts and mark their dot together (instead of one by one) while helping them and
making sure they understand the tool. This will save you time! Don’t use the real triads or diad as
examples since that will mess up the data; use a fictional example (like those proposed beneath).
4. After collecting the signification sheets with the triads and the diad, the facilitator (and not the
participants themselves) marks all participants’ dots on the triads and diad on the flipchart. This
will save time!
5. Make sure you construct the causal flow map in the direction of the participants so that they can
read/follow it. Remember that it is THEIR causal flow!
Tool 1 – Livelihood Matrix Analysis (50 min)
6. Start with a quick brainstorm to identify the main types of livelihood activities that exist in the
participants’ communities, related to roots and tubers (from inputs suppliers, production,
processors, traders, retailers/exporters).
-
Write the livelihood activities on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!). Drawings can be used instead of words if the participants are illiterate.
-
Cluster the cards that are similar or closely related, and give the clusters a name.
-
Keep asking for other types of livelihood activities related to R&T that exist in the
communities, until exhausted.
-
The note-taker writes the types of livelihood activities in the left column of the livelihoods
matrix in his note-taking template. The matrix looks as follows:
Livelihoods Matrix
Occupation
% of working
(examples
women in this
below)
livelihood
activity (now)
R&T farming
(specify)
R&T food
production
(specify)
Cassava lopping
(specify)
Transportation
(specify)
Trading
(specify)
…
(specify)
Small services
(specify)
OTHER not
R&T-based
livelihoods
% of working
women in this
livelihood
activity (before)
Rank according to
income PAST
YEAR
(highest at the top)
Score risk
(4=very high
to 1=very
low)
Major
changes in
past 5
years
7. Next ask each participant to estimate how many women / men were doing each type of livelihood
activity in the past year and before.
-
Place the cards underneath each other, and give every participants 10 stones.
-
Ask each of the participants to allocate their 10 stones to the cards, as an estimation of the
proportion of women / men that are doing each type of livelihood in their community in the
past year.
139
-
Let them do this one by one, and verify with each of them, if the total amount of stones s/he
allocated to all the R&T-based livelihood activities, all together, correctly reflects the
proportion of women / men involved in R&T (for instance 80% -or 8 out of 10). Ask her / him
if this proportion has changed in the past 5 years, and if yes, how much, and why. Do this for
every participant before moving to the next participant.
-
Don’t let participants use more than 10 stones! (if a particular activity should get less than 1
seed, then let the participant use a paper ball to indicate it’s less than one seed).
-
If other types of livelihood activities are mentioned during the exercise, add them to the list and
ask participants to re-allocate their seeds to include the new items.
-
If major differences occur in the participants’ allocations of stones to a particular livelihood
type, ask the participants why they are so different.
-
The note-taker takes carefully notes of important differences, total stone counts and average
percentages, and fills in the second column of the livelihoods matrix by adding all the stones
for each type of livelihood, dividing the total by the number of participants, and multiplying
this by 10.
8. Ask the participants now to rank the livelihood types according to income.
-
Start by asking: “In the past twelve months which types of livelihood activities typically had the
highest income?” Place this card high up on the ground.
-
Continue by asking: “Which had the lowest income? Place this card much lower than the other
one on the ground.
-
Ask the participants how this was different 2-3-5 years ago.
-
Next ask: “Which had the second highest income? And which the second lowest?” Place the
“second highest” card underneath the “highest” and the “second lowest” above the “lowest”.
-
Again, ask the participants how much this was different 2-3-5 years ago.
-
Repeat until all cards are given a place in the hierarchy.
9. Finally, ask the participants to score the R&T-related livelihood activities according to risk (scale
1-4; highest risk=4).
-
Ask: “How likely is there something to happen very bad to the R&T-related livelihood
activities?”
-
Ask them to describe the risks and consequences.
-
Those happening most often/likely with most severe consequences present the highest risk
(score 4).
10. Finally, ask the participants if any major changes happened over the past 5 years to the R&T-based
livelihoods, and which changes were the most significant ones, with the biggest impact for most of
the people in the communities, in particular those less well off.
-
Write the changes on MODERATION CARDS (no sticky notes!), one change per card, using
MARKERS (no pens!). Drawings can be used instead of words if the participants are illiterate.
-
Cluster the cards that are similar or closely related.
11. Output:
-
Livelihoods matrix properly filled in the note-taking template
-
Detailed notes on:
o the differences in proportional allocation of women / men to the R&T-based livelihood
activities, and the changes in past years in these proportions
o the risks threatening the R&T-based livelihoods in the communities
-
Cards with major changes that happened to the R&T-based livelihoods.
140
Tool 2 - Signification of livelihood changes (40 min)
12. Ask the participants to select the 2 most significant changes in the R&T-based livelihoods –one
positive and one negative.
13. Ask the participants to reflect on their own experience with one of these changes (2 mins):
-
Prompting question: “If you were invited by the local radio station to describe in no more than
2 minutes how you have experienced this change, what would you say?
-
Every participant is given 2 minutes to share his/her experience. In case the change doesn’t
relate to a participant’s livelihood, s/he can describe the experience of someone s/he knows
very well.
-
The note-taker carefully writes the individual experiences down on a uniquely coded
signification form for each participant, using her /his exact words (thus NOT interpreted or
summarized), and indicates whether the change/experience was positive or negative.
14. Explain the signification process.
-
Explain to the participants that the purpose of the exercise is to give a individually score their
own experience using the ‘diads’ and ‘triads’, and explain the tools (triads, diads) by means of
a fictional (!) example, such as:


“How did you spend your day yesterday? eating, working, sleeping”-triad
“What is your feeling about cooking? hate it, love it”- diad
-
Make sure that the people understand that the dot can be placed anywhere in the triangle or on
the diad.
-
Ensure that participants do their own scoring/dotting and not just copy the scores of their
neighbours.
15. Do the signification using the triads and diad.
-
Give each of the participants their own signification format, on which their experience was
written by the note-taker.
-
Read the first triad and ask the participants to place their dot in the first triad on year form,
using a MARKER (not a pen). They can select “N/A” if they the triad does not apply to their
experience.
The main effect of the change in my story is …
 Access to food
 Access to income
 Access to education
-
Continue with the second triad:
In my story the change was influenced by…
 Access to information
 Access to technology
 Access to finance
-
Continue with the third triad:
In my story the change was due to …
 My own efforts
 RTIMP (or the government)
 Other people or organizations
-
Continue with the fourth triad:
In my story the change was caused by…
 Production issues
141


-
Business relations
Processing issues
Conclude with the diad:
Based on my experience, I think…
 It’s worth to further invest in the production/processing of roots & tubers
 It’s worth to invest in the production/processing of new crops other than roots & tubers
16. Collect the dots and facilitate a quick discussion around the results as the basis for the causal flow
mapping.
-
Capture all the participants’ dots on the flipcharts and facilitate a brief discussion around the
pattern –e.g. by asking why most dots are in the same place, or why are they far from each
other?
17. Output:
-
Individual signification forms
-
Flipcharts with the collective scoring results on the 4 triads and the one diad (for the districtlevel sensemaking workshops)
-
Notes on the discussion around the results from the scoring
Tool 3 – Livelihood Change Causal Flow Mapping (60 min)
1. Explain to participants that now we will discuss further the causes and effects of these one-two
changes identified earlier.
-
Place the selected change statement in the middle of the flipchart – either in key word or a
symbolic drawing.
2. Probe first for the impacts of each change. Use different colours for indicating whether the impacts
are positive or negative.
-
Ask: “what has happened as a result of this change? how has it affected people’s lives?”
-
Write positive and negative impacts on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!) of two different colours, and place the cards on the right side of the
-
Probe for enough details (what, when, where, for whom, how many, why) to make it possible
to triangulate with other data (for instance from the household survey).
-
If someone mentions an indirect consequence, ask what happened before in order to link it to
the original change.
-
Probe for second and third line impacts by asking: “What has happened as a result of this
change/impact? And what has happened as a result of that?”
-
Continue probing until you have exhausted all impacts and have reached the level of “sufficient
access to food and income to lead and sustain a healthy and active life”
18. Next probe for the causes of each change in the middle of the flipchart. Asking for the causes of
these changes, and the causes of the causes (first, second and if needed also third line causes).
-
Ask questions like: “What has made these changes possible? What else contributed to or
influenced this change? What were the underlying causes of each of these direct causes?”
-
Write positive and negative causes on MODERATION CARDS (no sticky notes!) using
MARKERS (no pens!) of two different colours, and place the cards on the left side of the
-
Probe for enough details (what, when, where, for whom, how many) to make it possible to
verify with other sources. For instance, in case of ‘better infrastructure’ probe until you arrive
142
at something as specific as ‘the new bridge over the canal that links the village to the next
district and was built just before the latest elections’.
-
Keep on probing until exhausted and have reached the level of services and interventions.
19. Finally, ask the participants if all changes, causes and impacts are placed in the right order in
relation to each other and ask them to help you draw the connections.
20. Output:
-
Livelihood change causal change flow diagram (hard copy in the notes).
Closure (10 min)
2. Ask permission to take a photo of the diagram, share it with others and use it for discussion at the
district (together with other groups’ diagrams).
21. Output:
-
Photo of the causal change flow diagram.
-
List of people who want to participate in the district-level sensemaking workshop.
Description of the outcome for the participants of the exercise
143
E. SenseMaker Tools
Participant:
Change:
Positive/ Negative:
District:
Community Cluster:
Group Code:
Prompting question:
“If you were invited by the local radio station to describe in no more than two minutes how you have
experienced this change, what would you say?”
Experience:
Access to education
Access to Information
NA
NA
Access to
income
Access to
food
Access to
Finances
Access to
technology
Triad 2: In my story the change was
influenced by…
Triad 1:The main effect of the
change in my story is…
Business relations
Other people or organizations
NA
NA
My own
efforts
RTIMP
(Local Gov’t)
Triad 3:In my story the change was
due to…
Production
issues
Processing
issues
Triad 4:In my story the change was
caused by …
N
It’s worth to further invest in
the production/ processing of
roots & tubers
Diad 1: Based on my experience,
I think…
It’s worth to invest in the
production/processing of new
crops other than roots &
tubers production/
144
Annex 11: Constituent Feedback
A. What is Constituent Feedback (CF)
Constituent Feedback (CF)81is a lithe and low cost methodology for systematic listening to, and engaging in
dialogue with, key constituents in the development process as the basis for performance monitoring, giving
them real voice in defining, assessing, explaining and debating success. Feedback is best collected at key
points of a program delivery and during normal rhythms of implementation and monitoring.
While feedback can only be turned into voice through ongoing cycles of listening, analysing, engaging and
acting through the life of a program, it can also be used in participatory impact assessment and learning as
the basis for generating critical debate between sponsors, service providers and primary constituents around
impact-related issues. Combined with other methods, CF can help assess effectiveness and contribution to
impact of a program, by collecting perceptual data on the reach, quality and value/outcomes of services and
relationships developed.
There are many ways of collecting, analysing and reporting feedback data. Most essential is to select and
design the tools in a way that is appropriate to context and useful to all the constituents or stakeholders
involved –including the beneficiaries, service-providers and implementers, and the funders. This is in line
with the purposes and principles of PIALA.
B. When and how to use this method, for what purpose
Constituent Feedback is best used after the livelihood analysis at the end of the participatory research in
each locality. The purpose is to obtain quantified perceptual data on the reach, quality and outcomes of
specific program mechanisms for intended primary beneficiaries at key touch points of program delivery
and influence in the program’s ToC. Each CF session is built around three types of questions:



Reach and quality of services
Inclusiveness and quality of relationships (e.g. trust, voice, empowerment)
Value and outcomes of the services and relationships
For the evaluation of RTIMP, the key touch points are the most important links between the M’s and the
C’s in the ToC diagram that form the focus of evaluation. These are centred around the functioning of:



Farmer Field Forums (FFF)
District Stakeholder Fora (DSF)
Good Practice Centres (GPC)
In case of the DSF, also the contribution of the Supply Chain Facilitators to the functioning and outcomes
of the DSF are inquired. The MEF on the other hand is assessed in relation to and as an outcome of the
GPCs.
For each of these three program mechanisms, a focus group discussion is organised with the intended direct
beneficiaries around a small number of questions, which involves an individual and anonymous scoring by
each of the participant. The questions can ask for a simple YES/NO response, or HOW MANY, or they can
ask for a scoring on a scale. Adding few open questions helps generate more rich explanations and
contextual information.
81
Also called Constituent Voice (see www.keystoneaccountability.org).
145
Other constituents (or stakeholders) involved in these mechanisms will be surveyed on 3-4 related
performance questions, which they also will be asked to score. These scoring questions will be asked at the
end of each of the KIIs with any of the service-providers82.
C. With whom to use this method
This method will be applied with gender- and age-mixed groups sampled from the lists of participants or
primary target groups of the program mechanisms under evaluation. The groups need to be composed of
good representative mix of constituents targeted by the FFF, DSF and GPC in their catchment area. Ideally
the groups are composed of 10-12 people, of which 6-8 who actively participated in each of these
mechanisms, and the others not, half of which are women and half are men.
D. Time and materials
The total time needed to do the whole exercise is about one hour per CF session (including group reflection
& scoring). The materials needed include: CF facilitation questions and stones or seeds for the scoring.
E. Step-by-step guide
-
-
evaluation.
-
-
-
-
-
quotations.
-
82
Including: the FFF facilitators; the 7 research team leaders from CSIR, KNUST & UCC; the DDAs, BACs, SCFs and RTIMP
desk officers; the directors/leaders of the GPCs and other supply chain leaders, and of the PFIs.
146
-
-
-
30. Output:
-
Facilitation of CF discussion and scoring (40 min)
31. Arrange FGD participants in a circle. Participants should not sit behind each other, as they will then
be able see how other participants have scored.
32. The facilitator explains what the discussion will be about and how the participants will be asked to
score. The note-taker documents carefully takes notes of the qualitative feedback and indicates if
there is a major disagreement or difference for each of the questions, between whom, and why.
33. When the participants are doing the scoring of a question, the note-taker moves quickly around the
group to capture the scores and enter them into the Scoring Table.
34. The discussion and scoring should take no longer than 50 min per session.
CF Questions
DSF Questions:
a) Are you a member of the DSF? (Y/N)
b) How many DSF meetings have you attended past/two/three years ago? (numbers)
a) Out of ten, how many male/female/young farmers in your community participate in DSFs
past/two/three years ago? (number 1-10)
b) How useful have you found the information obtained through the DSF for improving your
business/livelihood? (scale)
 Why/why not? (open)
c) To what extent have you been able to obtain the technical support, training and services that you
needed past/two/three years ago? (scale)
 From the BACs?
 From the FFFs?
 From the GPCs
 Other from RTIMP
 Other -RTIMP?
d) How confident are you to express your views or ask for support/assistance at a DSF meeting?
(scale)
 Why/why not?
e) To what extent are decisions made at the DSF carried out in practice by all the actors? (scale)
f) Have you been able to buy/sell more past/two/three years ago through participating in the DSF?
(Y/N)
g) Have you been able to get a better price through the DSF? (Y/N)
 Why/why not?
h) Is the DSF catering to the needs of all those who need it? (Y/N)
 Why/why not?
i)
Are conclusions and actions coming out from the DSF shared in the wider communities? (Y/N)
147

Why/why not?
FFF Questions:
a) Is there a FFF in your community, or do you know of one in a neighbouring community? (Y/N)
b) How many FFFs have you attended past/two/three years ago? (number)
c) Out of ten, how many male/female/young farmers in your community participate in FFFs?
(number 1-10)
d) To what extent have you been able to obtain the technical support, training and services that
you needed from the FFF past/two/three years ago to expand your farming business? (scale)
e) Can anyone (including rich/poor, men/women, youth/elders) participate in the FFF and apply
what they learn? (Y/N)
f) How confident are you to express your needs and ask for help at a FFF meeting? (scale)
 Why/why not?
g) To what extent have you been able to apply what you learned at FFF? (scale – applied nothing --apply everything that I learn)
 Which technologies /which not?
h) To what extent did the training help you expand your farming business? (scale)
i)
Have you been able to produce more at lower cost past/two/three years ago through
participating in the FFF? (Y/N)
j)
Have you been able to sell more past/two/three years ago through participating in the FFF?
(Y/N)
k) Have you been able to obtain a better price through the FFF? (Y/N)
GPC Questions
a) Is there any GPC in your community, or do you know of any in a neighbouring community?
(Y/N)
b) How many GPC trainings/demos have you attended past/two/three years ago? (number)
c) Out of ten, how many male/female/young processors in your community are part of the GPC
and participate in these trainings/demo’s? (number 1-10)
d) Can anyone (including rich/poor, men/women, youth/elders) participate in the GPC and apply
what they learn? (Y/N)
 Why/why not?
e) Have you obtained any financing through the GPC to buy equipment and expand your
processing business? (Y/N)
f) Explain (which equipment? what kind of finance –e.g. loan, grant, gift? from which sources –
e.g. rural bank, commercial bank, family/friends, middle man, GPC?)
g) Have you heard of the MEF and understand what it is? (Y/N)
h) Have you or anyone else in your community applied for MEF? (Y/N)
 Why/why not?
i)
To what extent did the training, support and services help you to expand your processing
business? (scale)
148
j)
How confident are you to express your needs and ask for help at the GPC? (scale)
 Why/why not?
k) To what extent have you been able to apply what you learned at the GPC? (scale)
 Why/why not?
 Which technologies /which not?
l)
Have you been able to process more at lower cost past/two/three years ago through
participating in the GPC? (Y/N)
j)
Have you been able to sell more past/two/three years ago through participating in the GPC?
(Y/N)
m) Have you been able to obtain a better price through the GPC? (Y/N)
Closure (10 min)
149
Annex 12: Data collation and quality monitoring
A. What, When and Why
Data quality monitoring involves daily research team reflections on research processes and outcomes as the
basis for timely identifying data gaps/weaknesses and assessing the robustness of the evidence base being
built. To identify gaps and weaknesses, early data collation is needed. This is the process by which the
qualitative and quantitative data collected in each sampled district from both primary and secondary sources
is processed, integrated and linked to each of the causal claims in the ToC. Data quality monitoring thus
goes hand in hand with data collation.
The purpose of data collection and quality monitoring is:
(a) check on the robustness of the emerging evidence and identify areas where more data is required;
(b) enable data integration and linking; and
(c) reconstruct and compare actual causal flows with the program’s theory of change at a very early
stage during field research.
A set of questions for data quality monitoring and a basic tool for data collation is prepared for this. The
tool consists of a matrix to insert quantitative and qualitative data for each link in the ToC, and a simple
rating system for estimating: (a) the robustness of the emerging evidence; (b) the extent and scope of each
causal link; (c) the extent and quality of program contributions; and (d) the interference of other influences.
The average score for each contribution claim in the ToC obtained from the scoring (0-6) of each of its
links on these four criteria, will provide a total value from ‘highly unsatisfactory’ up to ‘highly
satisfactory’83.
B. With Whom
The researcher who leads the research in a particular district should conduct the data collation and
reflection together with his/her research assistants.
C. Time & planning
Every evening the two teams working in parallel in the same district must come together for two-three
hours to critically reflect on the processes and data obtained that day, and do the data collation together.
The exact time needed will depend on early data entry and processing of the quantitative data. At a
minimum however, 2 hours will be needed to:



discuss the data and reflect on the overarching evaluation/learning questions in relation to the ToC;
discuss observations on the process and participation and power issues that may have influenced
the data and generated potential bias;
fill in the data collation table, using the raw data sheets from the field and preliminary finding from
quantitative data entry and analysis.
While reflections on the research process will be mostly important in the first two days, focus should shift
towards filling in the data collation table and ensuring sufficient data is obtained to research to conclusions
from the third day onwards. Although data collation should start as soon as the first raw data sheets are
coming in, so basically from day one, findings from quantitative analysis likely will become available only
by the third or fourth day, so mid-way the research process in a district. Hence sufficient time should be
83
These are similar to the rating values used by IFAD in supervisions and reviews.
150
booked with the entire team in particular on the fourth and fifth evening to do the data collation and take
stock of the evidence base.
D. Questions for data quality monitoring
The following set of questions should be used to guide the evening reflections on the research processes
and outcomes, as the basis for ongoing data quality monitoring and estimating the robustness of the
evidence that is built from that.
It is of crucial importance to regularly (if not daily) reflect on these questions to make sure all criteria are
met by the time the field research in a locality or district is running towards its end. Major data gaps or
weaknesses MUST be identified and reported in time, which is before the sensemaking workshop takes
place in that locality/district.
1.
Were we able to conduct the interviews and FGDs as planned and in a systematic manner?
 If not why not? What were the constraints? What do we need to do differently? What do we need
to redo?
 Did we consult with key informants on how local institutions and processes are called/recognized
by people to help us formulate our questions in a way people can understand what we're asking
for?
 Did we keep sufficient focus and pace in the focus groups?
 Did we ask the right questions to obtain the right data?
 Did we follow the right sequence of tools?
 Did we prompt and probe well enough to obtain sufficient detail in our data to enable us to verify
it with other sources?
2.
How well did we facilitate discussions to ensure active engagement and equal voice?
 How actively were people participating? Who was silent? Who was dominant?
 How were differences in views/opinions in the group discussions addressed and recorded?
 What hindered us from obtaining honest responses?
 What can be improved in our facilitation approach as to overcome these difficulties and improve
the quality of participation?
 How well did we facilitate the discussions to ensure sufficient focus? Did we probe well and go
beyond initial superficial answers? What can be improved?
3.
Were we able to properly code and document all data?
(including descriptive data from the brainstorms, explanatory data from the causal flow mappings and
interviews, and evaluative data from the scorings)
 Are all raw data sheets properly coded?
 Have scores been properly documented?
 Have qualitative explanations and differences in views and perceptions been properly documented
with sufficient detail and accuracy?
 Have we accurately captured people’s views in their own words?
 Have we taken clearly visible photos of the diagrams?
 Are notes and flipcharts legible?
 Have we properly documented our process observations?
 Are all raw notes accessible in hard & soft copy?
4.
Were we able to obtain robust and sufficient data on the causal links in the ToC?
 Did the interviews and focus group discussions provide us with the information that is needed to
answer the evaluation/learning questions and put the assumptions at a test?
 Is the information sufficiently inclusive of different perspectives, dependable and consistent?
 Is the data we collected sufficient to estimate with confidence the extent and scope of the causal
link in the ToC it was supposed to inquire?
 Do we have sufficient detail to support our conclusions? If not, how can we obtain the level of
detail that we need to make our conclusions hard?
151


5.
Have we adequately dealt with positive respondent bias?
Have we adequately dealt with our own researcher bias?
Were we able to collect sufficient data on the value and reach of specific program mechanisms?
 Is the data we collected sufficient to estimate the extent and quality of RTIMP influence in the Mlinks?
 Have we sufficiently inquired other influences on the changes and impacts observed?
 Were we able to sufficiently cross-check our data with different sources?
 Were we able to inquire those not reached by the program to assess the reach and value of specific
program mechanisms?
 Do we have enough data on the different with/without configurations of program mechanisms to
capture variability in program treatment and impact across the program area or country?
E. Data collation table
The table on the next page serves as a basic tool for the systematic integration/collation and analysis of the
data collected in every district from primary and secondary sources related to each link in each causal claim
of RTIMP’s ToC.
For each causal link, evidence obtained is bundled in the 2nd–4th columns, which should permit to identify
data gaps and draw conclusions. In the 5th column, the robustness of evidence collected on each particular
link is assessed. If the evidence is sufficient, conclusions can be drawn on the extent and scope of each
causal link, and the extent and quality of RTIMP’s contributions on each of these links in each contribution
claim, taking also into account other influences. The average score for each contribution claim obtained
from the scoring (0-6) of each of its links provides a total value from ‘highly unsatisfactory’ up to ‘highly
satisfactory’ similar to the rating values used by IFAD in supervisions and reviews. The rating values are
described in the table below.
Contribution claim
6
Highly satisfactory
5
Satisfactory
4-3
Moderately
satisfactory /
unsatisfactory
2
Unsatisfactory
1
Highly unsatisfactory
0
Insufficient evidence
Quality descriptors for the contribution claim
 Evidence on most causal links is robust.
 Extent and scope of all CMCOI- links in the contribution claim is convincing.
 RTIMP interventions and achievements can be demonstrably linked to the changes.
 Negative influences (intended or unintended) minimized, and
positive (intended or unintended) influences changes maximized.
 Evidence on causal links is quite strong, although some gaps still exist.
 Extent and scope of most CMCOI-links in the contribution claim is convincing.
 RTIMP interventions and achievements can be demonstrably linked to most of the
changes but perhaps not to all.
 Some negative influences might have occurred that were not entirely mitigated.
 Evidence on some causal links is generally strong but on others remain rather unclear
and inconsistent.
 Extent and scope of some CMCOI-links in the contribution claim is not entirely clear or
not entirely convincing.
 Although there is evidence of RTIMP contributions to some changes, the evidence is
insufficient and lacking on others.
 Unintended influences that affected the results not entirely mitigated.
 Evidence on causal links is rather weak and/or inconsistent.
 Extent and scope of most CMCOI-links is unclear and/or unconvincing.
 Evidence of RTIMP contributions to most changes is lacking or insufficient.
 Unintended influences have been largely overlooked.
 Evidence on causal links is negative.
 Extent and scope of CMCOI-links is limited.
 Evidence of RTIMP contributions to all changes is lacking, insufficient and/or negative.
 Other influences have been neglected, indicating mismanagement.
 There is insufficient information available on changes or causal links
to assess the contribution claim.
152
Data Collation Table:
Please empty the spaces in the three middle data columns and fill in all the data there. Draw conclusions and score the strenth of evidence for each link only
when the data is complete for that row. Calculate to overall score for the impact claim and each of the contribution claims only when the conclusions and
rating for each of its causal links are completed.
Causal link
Secondary data
Primary QUANT data
IMPACT CLAIM – POVERTY REDUCTION
I2→I1
60-90 gender/age 2010 Ghana Living
disaggregated
Standard Survey
household survey
report
O3+O2+O1→I2  RTIMP RIMS baseline
and other M&E data
Data Gaps:
Primary QUAL data
Strength of evidence
Conclusions
(Score 0-6, whereby 0 means
there is insufficient
information, and provide
critical notes on remaining
data gaps and process
observations suggesting
possible bias in the data)
(Score 0-6, whereby 0
means there is insufficient
main conclusions about: (a)
the extent and scope of each
causal link; (b) the extent
and quality of RTIMP
influence in the M-links;
and (d) interference of other
influences at all levels)
4-6 sessions of generic
change analysis & causal
flow mapping in
gender/age-specific focus
groups
Total average score
(0-6):
153
Causal link
Secondary data
Primary QUANT data
Primary QUAL data
CONTRIBUTION CLAIM OF COMPONENT 1 – ENHANCED MARKET LINKING
M1c+M1b+O2
4-6 sessions of livelihood analysis & causal flow
 DDA reports
+O3→C1b
mapping in gender/age-specific focus groups
 RTIMP Enterprise
C1a+(M1)→O1
KIIs with DDAs, BACs,
Record Books (ERBs)
C1b+M1a→C1a  ZOCs progress
SCFs, GPCs and other
supply chain leaders
reports
(SMEs, aggregators and
 MoFA and DADU
exporters) and off-takers
Organisational
(industries, food
Capacity Assessments
traders…)
 RTIMP M&E data
2-3
constituent
feedback
sessions
with mixed groups
(including 2014
of (non-)DSF participants
thematic impact
studies on DSF & SCF
and IEC)
Data Gaps:
Conclusions
11.6.1.1.1.1.1
Total average score
(0-6):
154
Causal link
Secondary data
Primary QUANT data
Primary QUAL data
CONTRIBUTION CLAIM OF COMPONENT 2 –ENHANCED R&T PRODUCTION
C2a+C2b→O2
 Review of RTIMP
productivity surveys
M2a+M2b+(M2c)
KIIs with FFF
and progress reports
+M1c→C2a
facilitators, extension
from the SRID, GLDB,
M2c→C2b
agents, DDAs, DADU
DDAs and ZOCs
officers, and 7 research
 RTIMP M&E data,
team leaders84 of the
(including the 2014
regional research
thematic impact
institutes (CSIR, KNUST
assessment of FFFs)
& UCC)
2-3 constituent feedback sessions with mixed groups
of (non-)FFF participants
Data Gaps:
84
Conclusions
causal link; (b) the extent and
quality of RTIMP influence in
the M-links; and (d)
interference of other
Total average score
(0-6):
There are 7 research team leaders, involved in the FFFs, 5 of which are based in Kumasi, 1 in Cape Coast, and 1 in Tamale.
155
Causal link
Secondary data
Primary QUANT data
Primary QUAL data
CONTRIBUTION CLAIM OF COMPONENT 3 – ENHANCED R&T PROCESSING
M3b→
 RTIMP and REP
C3a+C3b→O3
M&E data and
M3b+M3c+C1a→
KIIs with GPCs, BACs and
supervision reports
C3c
PFI local branches
(including the 2014
thematic impact
2-3 constituent feedback sessions with mixed groups of
studies on MEF and
(non-)GPC participants (including MEF beneficiaries)
GPC)
 IFAD/FAO 2014 study
on matching grant
facilities in Ghana
Data Gaps:
Conclusions
Total average score
(0-6):
156
Annex 13: Participatory sensemaking at the district levels
A. What, When and Why
The purpose of participatory sensemaking is to enhance the empowering value of impact evaluation by creating
the opportunity for program stakeholders to validate, challenge and strengthen the evidence during the research.
At the local level, this is done by instantly processing the data collected during fieldwork, and presenting back
rising evidence and remaining data weaknesses to local research participants in a small sensemaking workshop.
Doing this before finalising fieldwork in every locality helps improve and strengthen the evidence base, while
avoiding top-down data extraction and researcher-dominated analysis. By stimulating critical reflection and
dialogue around the emerging evidence between service-providers, beneficiaries and other local actors involved
in the program, an additional layer of meaning or explanation is generated of why certain program mechanisms
worked well (or not so well) in certain locations, and had a significant and sustainable impact on rural
livelihoods and poverty (or not).
The local sensemaking process can best be organised in a small workshop at the level of analysis above the
household level. In some cases, this will be at the level of the village; in other cases (and particularly in IFADfunded value chain programs), this will be at the administrative or organisational level of the market system,
which for RTIMP is at the district level.
The objectives are to




discuss and validate the emerging evidence on the major changes that have occurred in rural poor
people’s livelihoods in the past 5-7 years, and the causes of these changes;
obtain additional information on issues that remain unclear and fill in data gaps;
create an opportunity to collectively reflect on the livelihood opportunities and constraints of the R&T
commodity chains for coming generations;
reflect on the value of local stakeholders’ participation in the impact evaluation for improving their
development practice.
B. With Whom
The participants in the local sensemaking workshop should form a representative sample of the local
stakeholders who participated in the research. Since the majority of the research participants are the primary
beneficiaries, they will form the largest group in the workshop. In the case of RTIMP these are the small R&T
farmers, seed producers and processors. Furthermore also district officials, service providers and supply chain
leaders in the RTIMP will be invited. The total amount of participants is estimated at 30-40 participants.
A provisional list of the participants in the district sensemaking workshops of the RTIMP evaluation is as
follows:




7officials (including the District Head, the DDA, the extension officer, the BAC officer, the RTIMP
officer, the FFF facilitators, and the research team leader);
10-12 processors (min 50% women and 50% <age 35)
10-12 farmers (min 50% women and 50% <age 35)
8 other supply chain actors (including supply chain leaders, SCFs, local PFI branches, local off-takers,
transporters, etc.)
157
C. Time and Materials
Preparation of the workshop is best arranged for at the start of the field research in a locality or district, so that
local leaders can timely inform people and start preparing the logistics. At least half a day after finalising
fieldwork (including data collation) is needed to prepare the inputs for the workshop. The workshop itself will
take average 4 hours but will vary depending on the number of participants. The local sensemaking workshop
for the RTIMP evaluation will be held at the district level and expect to host around 30-40 people.
Printing facilities and workshop materials need to be at hand. The venue needs to be big enough to host 40
people (with average two square meters per person) and have sufficient tables (6) and chairs (50) for small group
work, and walls on which to hang the flipcharts.
If the workshop has to be held in an open village space, then rope and nails or clothespins can be used to hang
the flipcharts for instance on/between trees, or they can by laid out on the ground with stones to keep them in
place. Cardboards can be used as an improvised table to put the flipcharts on for writing.
Preparation of venue and workshop inputs (4 hours)
1. Prepare maximum 10-12 statements that reflect the evidence on each of the links in the ToC diagram
(e.g. 2 related to the impact claim; 2 for the first, 2 for the second, and 2 for the third contribution claim;
and 2-4 extra related to those links where evidence is weak).
The statements must reflect the evidence on causal links (instead of the individual changes or causes). It
must do so by either confirming or contesting the link. The purpose is to (a) validate evidence and (b) fill
data gaps.
Examples of ‘evidence statements’ from an evaluation in Vietnam in 2013 are:






Many technical trainings and study-tours for new cultivation technique such as durian planting,
pig and goat raising have been organized in past years by the Farmer Union, but they haven’t
caused any income increase for the poor and the near-poor.
New credit (other than from the Social Policy Bank) has been offered to the poor and the nearpoor in recent years by the WU and the Agribank, but the preconditions of these credits are too
difficult for them to accept the offers.
There are no trust-groups in this community/area that accepted a loan from the Women or
Farmer Union.
Cows, goats, rice, and other goods are sold at lower prices to middlemen, because there is no
local support mechanism to help fixing the prices.
Cow raising has increased because poor people now can borrow money from the Agri-bank.
By buying and raising a cow with their loan, people can invest in the education of their children
and escape from poverty,
2. Make a flipchart with the evidence statement as follows:
Statements of findings Scores: true / kind of true / untrue
Statement 1
…
Statement 10
158
3. Also make a flipchart with PIALA statements for assessing the value of participation in the impact
evaluation with the participants:
PIALA statements
Scores: true / kind of true / untrue
1. The purpose of the research and this workshop
are still unclear to me.
2. The discussions during the research and the
workshop have not given me any new insights or
power to change things.
3. It was hard for me to understand the discussions,
which made it difficult to participate.
4. I’m not used to participate in discussions like
these, which made me feel uncomfortable.
5. The researchers asked too many questions and
left not enough time to us to discuss and analyse the
issues ourselves.
4. Print sufficient hard copies of the sensemaking micro-survey.
5. Print or write the PIALA and the evidence statements on A4 sheets (1 statement per sheet), and prepare
the visuals that must support the evidence statements on flipcharts:
Clusters and Links in the ToC
Qualitative visuals
Quantitative evidence
Impact claim:
 4-6 generic change
 Triads 1& 4
I2→I1
diagrams (if available on
 1-2 quant bars from HH
O3+O2+O1→I2
flipchart85)
survey
Contribution claim 1
 4-6 livelihood change
 Diad 1
- Market linking:
diagrams (if available on
 1-2 quant bars from the CF
M1c+M1b+O2 +O3→C1b
flipchart)
with DSF participants
C1a+(M1)→O1
C1b+M1a→C1a
 Triads 3 & 2
- Enhanced production:
M2a+M2b+(M2c)+M1c→C2a
with FFF participants
M2c→C2b
 Triads 3 & 2
6. M- Enhanced processing:
a M3b→C3a+C3b→O3
with GPC participants
kM3b+M3c+C1a→C3c
e
sure you get access to the venue an hour before the workshop starts, to arrange the space:




Put 4 tables in the opposite corners of the room where the group discussions will take place.
Remove all other tables, and keep only chairs, as to create space.
Hang the flipcharts with the evidence visuals on the walls.
Hang at the front of the room the two flipcharts you should prepare in advance: one with the
statements on the findings, and one with the PIALA assessment statements. See below:
Workshop flow (4 hours)
INTRO (15 min):
85
If it was not possible to do it on flipchart and thus you only have pictures of the causal flows done with natural materials in the
communities, then obviously you won’t be able to show these. In that case just work with what you have: the statements and the quant
bars. If you have time left for preparation you can try to copy some of the flow diagrams from your pictures on flipcharts.
159
7. Make sure that all participants find a comfortable place to seat. Open the meeting by welcoming
everybody and explaining the purpose and the expected total duration of the workshop.
8. Explain that the outcomes of this workshop will be presented at a bigger national workshop where also
all other districts will present their results, and that some of today’s participants will be invited to
participate in this bigger event. All communities that participated in the research are invited to attend the
district-level and national workshops. Explain that most participants in the research are like them: small
R&T farmers, seed growers and processors. Explain that their participation is considered crucial for the
improvement of the R&T sector.
9. Divide the participants into 4 groups (farmers, processors, other chain actors, and officials) and divide
the researchers among the groups for the facilitation. Note that the farmers group and the processors
group will need stronger facilitation to encourage active debate and equal voice, since these are the
biggest groups with most illiterate and mixed women and men, younger and older people.
VALIDATION OF EVIDENCE (150 min):
10. Facilitate discussion around the 10 statements in the 4 groups (max 90 min; 1 researcher/facilitator per
group) as follows:








Explain the purpose of the exercise: Since this research is about changes in their lives and what
caused these changes, we want to give back to them what we found, and would like them to discuss
and improve our findings.
Start with briefly presenting the quantitative evidence for the impact claim (including quant bars
from the household survey, and triads 1 and 4). Ask the participants if they think this is a fair
representation of the major changes in their lives in the past 5 years.
Next read the first 2 statements related to the impact claim, and ask the participants what they think
about this statement. Fuel the discussion, if requested by the group then amend the statement, and
finally ask the participants to score it “true”, “untrue”, or “kind of true/untrue”. If rated it in the
middle, then ask if they think it’s in the middle but not true, or rather in the middle but not wrong,
and ask why. Indicate the rating on the back of the card and carefully note down the reasons why. If
amendments are requested, make then with a red marker in the statements.
Continue with briefly presenting the quantitative evidence for the first contribution claim
“market linking” (including the diad and 1-2 quant bars from the constituent feedback data). Ask
the participants if they think this is a fair representation of the major changes in the R&T business
in the past 5 years.
Next read the 2 statements related to this first contribution claim, and facilitate the discussion and
the rating in the same way as for the impact claim.
Do the same for contribution claims 2 and 3. As the participants are getting more comfortable,
speed up the pace.
Close the discussion by appointing a participant who will present the results from their discussion in
plenary.
Carefully make notes of the process in your notebook: who spoke, who didn’t, who made the final
decision, how long did it take to reach agreement…
11. Invite the groups back into plenary for presenting and discussing their amendments and ratings with the
larger group (max 60 min).

The facilitator first reads each statement out loud. Next the farmers and processors groups present
their amendments and ratings. Subsequently the supply chain actors and officials are asked to add
theirs.

A brief discussion is facilitated around the differences in the ratings, and reference is made to the
flipcharts presenting the original data. ONLY DISCUSS THE DIFFERENCES; BUT DON’T TRY
160
TO REACH AGREEMENT. Most important is that the participants understand the differences in
views and perspectives, and that the views of the poor HHs are considered central. Carefully take
notes of these differences.

Finally, ask the participants what they could do themselves to change the situation in their
community, based on their analysis, and what is preventing them from doing this.
METHODOLOGICAL REFLECTION (60 min):
12. Let the participants return to their 4 groups (farmers, processors, other chain actors, and officials) and
facilitate discussion around the 5 PIALA statements in these 4 groups (max 40 min; 1
researcher/facilitator per group),following a similar process as for the evidence statements but faster
since there are no visuals that need to be explained and discussed.
13. Then invite the groups back into plenary for presenting and discussing the results, again following the
same process as for the evidence statements (max 20min).
CLOSURE (15 min):
14. Ask the participants to individually fill in the micro-survey about the village feedback.
15. Close the village meeting by thanking the participants for their participation. INVITE THE
PARTICIPANTS FOR A CLOSING DRINK.
161
Annex 14: Participatory sensemaking at the national level
A. Objectives and outcomes
At this stage of analysing and understanding the large amount of evidence that has been produced, both national
and local stakeholders are invited to engage in a sense-making process to:
 validate observed changes and impacts across districts,
 discuss possible explanations,
 value RTIMP’s contributions among other influences, and
 reflect on the evaluation approach (PIALA).
The concrete expected outcomes are:
 an approved set of changes and impacts as was envisioned by RTIMP,
 an approved set of actual changes and impacts that have occurred and were most significant in the
Northern, Middle and Southern zones respectively,
 an approved set of overlaps and differences between actual and envisioned changes and impacts,
 an approved set of explanations for these overlaps/differences,
 an appreciation of RTIMP contributions to the actual changes and impacts,
 feedback on the evaluation approach, and its possible use in future programming.
B. Participants
About 100 invitations were sent to ensure a diverse set of views are captured, representing the different
stakeholders at national, zonal, district and community levels. These stakeholders know about the RTIMP and
take an active interest in the outcomes of this evaluation.
Broadly these include:
 At the national and zonal levels (23 participants):
o MoFA (4)
o RTIMP (10)
o GASIP (2)
o IFAD (4)
o Participating Financial Institutions (4)
o Research Leaders (4)
o Supply chain facilitators (4)
o Other stakeholders (4)
 At the district level (21 participants):
o District officials (12)
o Business Advisory Centers (4)
o GPC leaders, Off-takers and Supply chain leaders (9)
 From the community clusters (45 participants):
o R&T farmers and processors (30)
o R&T processors (15)
162
C. Design principles
The sense-making model builds off following design principles:
1. creating space for equal and meaningful participation;
2. focusing on changes broader than intended results;
3. prompting honest critique and debate between different stakeholder perspectives on these changes;
4. using visuals and statements as a mechanism for presenting the evidence of causes and impacts;
5. building the analytical chain of actual changes;
6. mirroring the actual with the envisioned changes in the ToC for probing its causal links and
assumptions;
7. developing an aggregated perspective despite diversity;
8. developing a replicable model (in terms of flow and design principles) for participatory sensemaking.
163
D. Straw Dog
Day 1 (6th May 2015): Connecting People - Clarifying the Vision and Checking the Condition – Identifying the Explanations
Time
Mod
Topic
Process
Preparation - Materials - Translation
8:30 - 9.00
1a
Introduction
 Brief words of
welcome
 Administrative
issues
Participants receive a folder containing
workshop documents and a colored card
with the opening assignment upon
registration
Participant workshop pack:
 hand-out with opening assignment (assigned color)
 hand-out with workshop objectives & agenda
 hand-out with generic diagram
 copy of presentation on the impact evaluation
 hand-out with “check-the-vision” or “check-thecondition” assignment (assigned color& subgroup
code)
.
Person
Responsible
Glowen
Colors (Mod 1a & 3b): green for farming, blue for
business relations & markets, purple for processing,
and red for lives & livelihoods.
9.00-9.15
1b
Opening
assignment
15 min. Participants are welcomed and
asked to form a group of people with 3
different color cards and share their
changes
Opening assignment:
Essence of this evaluation is not about RTIMP, but
about building on RTIMP, learning about the effort
that have been put into resource-poor farmers to do
better in the next GASIP. It’s not about government
bringing more money, but it’s about government better
reaching the majority farmers and engaging them in
the planning, implementation, evaluation and learning
of the activities affecting their lives and livelihoods.
Essence of this research is to learn about the changes
people have experienced in their livelihoods, better
reach the majority of the farmers, make better policy
for the future,.
Glowen
164
9.15 - 9.30
1c
9.30 - 9.50
1d
9:50 -10:15
2
10:15-10:30
3a
Welcome by
RTIMP
Workshop Purpose
 Why we are here
 Objectives
 Agenda
 PIALA
presentation
 RTIMP Theory
of Change
 Rules and
expectations
Evaluation
Findings
Opening words by Mr Minka Fordjuor,
Regional Director for Agriculture
Link opening assignment to workshop
purpose and structure by presenting the
RTIMP ToC as a tool for assessing changes
in R&T livelihoods and the causes and
contributions to these changes
Checking the
vision & condition
 validation of
vision
statements
 validation of
evidence
statements
30 min. Explain the task and make the
groups.
 Divide in 4 cluster groups of 25
participants, each guided by
1.5 facilitator
 Each group is assigned to work on a
particular cluster in the ToC:
1. improved lives & livelihoods
2. enhanced market-linking
3. enhanced farming
4. enhanced processing
 Split each of these 4 groups into 4
teams of 6 participants:
o 1 team is asked to discuss and
validate the vision statements
(Check the vision assignment)
o the other 3 teams are asked to do
the same with the evidence
statements (Check the condition
assignment), but for the Northern,
Posters with generic diagram (10)
Mr
Agyekum
Glowen
PIALA: Approach &, design
PIALA, the design of the impact
evaluation, and the field research conducted
in past months
Brief presentation of the major findings and
draft conclusions to be given in a statement
format.
Essi &
Anthony
 posters with vision diagram (1 per group)
 posters with condition diagram (1 per group)
 posters with vision clusters (1 per group)
 posters with condition clusters (3 per group)
 posters with sensemaker results (2 per group)
 posters with household survey results (2 per group)
Nana and
her
facilitation
army
(Adinda
process
facilitator)
 vision statements per cluster
 descriptive evidence statements per cluster
 RTIMP logframe (1 per group for the vision team)
 10 copies of the sub-report with sensemaker
findings (for the lives & livelihoods group)
 10 copies of the sub-report with household survey
findings (for the lives & livelihoods group)
 document with production & processing triads (2 &
165
Middle and the Southern zone
respectively
10.30 -11.00
11.00 -12.30
Coffee Break
3b
Checking the
vision & condition
continued
90 min. Discussion & validation of
statements in small teams.
Vision teams that finish earlier spread
themselves over the Northern, Southern and
Middle condition teams to listen only.
3) for the “farming” and “processing” clusters
 document with the diad for the “market-linking”
cluster
Colors:
green for farming, blue for market-linking, purple for
processing, red for lives & livelihoods.
Participants are invited to take a cup of coffee while
moving into their groups and reading their assignment
Check-the-vision assignment:
Read all the vision statements, and discuss and decide
with your team if these reflect the changes that RTIMP
had envisioned.
Make amendments where needed.
For each statement, check the logframe targets and
baseline figures, and decide on a realistic figure of
achievement.
Last, past the amended vision statements (with figures)
onto your cluster diagram.
If your team finishes earlier then the other teams, then
join of the condition teams (Northern, Southern or
Middle Zone) to listen only –NOT TO SPEAK!
Nana and
her
facilitation
army
(Adinda
process
facilitator)
Check-the-condition assignment:
Read the evidence statements, one by one, and decide
if it is:
 true,
 kind of true, or
 untrue.
Write at the back-side of the statements WHY you have
scores it as such.
Read all the statements that are “true” or “kind of
true” over again, and discuss and decide if all together
constitute what has changed for your cluster in your
region. If anything is missing, please write it down on
a moderation card.
166
Last, paste the statements with the ratings onto your
cluster diagram.
12.30 -13.00
4a
Mirroring
envisioned with
actual changes
90 min. The vision team presents their
cluster diagram to the Northern, Southern
and Middle condition teams, who then one
by one present their cluster diagrams
against the vision.
Nana and
her
facilitation
army
(Adinda
process
facilitator)
13.00 -14.00
14.00 -15.00
15.00 -16:00
Lunch
4b
5
Mirroring
envisioned with
actual changes
continued
Reconstructing the
causal flow
1ST ITERATION OF
CONTRIBUTION
ANALYSIS
The facilitators start to collect their teams at 13:50 to
make sure that Mod 4 is finished in time (by 15:00)
The teams discuss together the extent to
which envisioned changes have been
achieved in their cluster, and identify:
 the key areas of divergence: where
there is low, no, or unanticipated
change
 the key areas of difference compared to
the other zones: where there is
significantly more or less change
The 4 teams formulate their final
conclusions on flipcharts to present to their
peer teams in the other groups.
15 min. Explain the task and reorganize:
 Let move people into four groups, each
guided by 1.5 facilitator:
1. group with Northern teams
2. group with Southern teams
3. group with Middle teams
4. group with Vision teams
 The teams one by one put their poster on
the wall, making it fit while presenting
its content to the others –first “lives &
livelihoods” team starts, next the
“market-linking” team, next the
Nana and
her
facilitation
army
(Adinda
process
facilitator)
4 empty walls where 4 posters can be pasted to
reconstruct the entire RTIMP causal flow diagram
Nana and
her
facilitation
army
(Adinda
process
facilitator)
167
16.00 -16.15
16.15– 17.30
Coffee Break
6
Treasure hunt for
change
“farming” team, and last the
“processing” team.
 The groups then discuss the deviations
from the vision and the differences with
the other zones, and identify the main
reasons & explanations.
 All groups formulate their final
conclusions on flipcharts to present to
the other groups, and appoint 2
rapporteurs
While grabbing a coffee, the participants
are askedto move with their group
clockwise to the space of the next group.
75 min. Gallery walk and treasure hunt
(15 min per group + 5min transition time
between groups)
Participants are tasked to, individually,
identify the one area of change that needs
more effort in the next GASIP –and
formulate concrete ideas for this prolonged
effort on a colored card (the color
indicating the area: farming, processing,
market-linking, lives & livelihoods). The
ideas need to be prepared for the opening
discussion on the second day, so
participants are invited to further reflect and
think of good ideas over and after diner.
Nana and
her
facilitation
army
(Adinda
process
facilitator)
168
Day 2 (7th May 2015): Connecting People - Clarifying the Vision and Checking the Condition – Identifying the Explanations
Time
Mod
Topic
Process
8:30 - 9.30
7
Opening
discussion
15 min. Welcome, agenda
9:30 - 10:45
8
Valuing RTIMP
contributions
among other
influences
2ND ITERATION OF
CONTRIBUTION
ANALYSIS
45 min. Fishbowl discussion in which the
participants share and discuss their ideas for
prolonged effort on key areas of change
where RTIMP has not been very successful.
15 min. Split into groups and teams, and
per team select 6 most important
explanatory evidence statements.
Preparation - Materials - Translation
Person
Responsible
Glowen
Structure of fishbowl discussion to ensure equal
participation of farmers and processors
 2 empty flipcharts with 4 sliding scale
(1 for each cluster)
60 min. Discussion and rating of RTIMP
contribution in teams
 posters with generic diagram (1 per group)
 posters with the CF Quant Bars (2 per group)
Process:
 3 local groups organized per zone
(North, South, Middle) plus 1 national
group, each guided by 1 facilitator
 Each group goes to the break-out space
near the wall with the reconstructed
causal flow of their particular zone (cf.
mod 5)
 Each groups select all the evidence
statements they totally agree with, look
at the explanatory evidence of each
statement (at the back side of the
evidence statements)
o compare with the vision
statements, and
o rate RTIMP contribution in a
canvas that has 2 sliding scales:
o From positive to negative
o From strong to totally overrun
 Zonal posters with constituent charts (2 per zonal
group)
 Explanatory evidence statements per cluster
Nana and
her
facilitation
army
(Adinda
process
facilitator)
169
by other influences (with strong
with other influences”, and
“weak with other influences”
in between)
o Number the statements and put the
numbers into the canvas
Max. 10min per statement (60min)
10:45 -11:00
11:00 -12:00
9
12:00-12:30
10
12:30 -13:30
13:30 -15:30
15:30 -16:00
11
11
Coffee/tea break
Plenary sharing
and discussion
Micro-survey
Prioritization of
PIALA topics
Lunch break
PIALA reflection
Formal closing by
IFAD
One group after the other presents in
plenary
Mini-survey on PIALA is filled in.
Glowen
Meanwhile prioritization with national
stakeholders of PIALA aspects that need to
be discussed in the afternoon
Adinda &
Glowen
Discussion with national stakeholders of
priority aspects related to PIALA
Nana and
her
facilitation
army
Words of thanks and looking forward by
IFAD Country Program Manager, Mr. Ulac
Demirag
PDA team
(Adinda
process
facilitator)
Mr. Ulac
Demirag
170
Annex 15: Fieldwork Schedule
Scenario 1: Four days with only 1 cluster
Day
Location
Day 1 District Level
and Community
Level
Time
early
morning
Team A
"District Level entry",
organization of 5-6 KIIs and
preparation for sense-makingworkshop
Team B
community entry in 1st
community and 4 household
surveys
Morning
KIIs: DDAs and BAC
KIIs: Supply Chain Leaders
(GPCs and SMEs)
Afternoon
FGD: Generic Change
Analysis, men
KIIs: Supply Chain Leader
(Retailers) and PFIs (local
branches)
FGD: Generic Change
Analysis, women & youth
6 household surveys in 1st
community
Debrief and data entry
community entry in second
surveys in 2nd community of
cluster
FGD: Livelihood Analysis,
women & youth
community entry in second
surveys in 2nd community of
cluster
men
afternoon
FGD: Constituent Feedback,
participants in DSF /GPC
FFF
late
afternoon
third community entry/chief
visit and 5 household surveys
in 3rd community of cluster
third community entry /chief
visit and 5 household surveys
in 3rd community of cluster
Evening
early
morning
Morning
afternoon
late
afternoon
Evening
early
morning
Morning
afternoon
late
afternoon
Evening
mopping up
late
afternoon
Day 2 Community Data
Collection
Evening
early
morning
Morning
Day 3 "District Level"
Day 4 "District Level"
Preparation of sensemaking-workshop
Mopping up
Mopping up
Half-day of sense-making workshop at district level
Travel to next District
Debrief and Data Entry
171
Scenario 2: Five-six days with 2 clusters
Day
Location
Day 1 District level
and First
Cluster
Time
early
morning
Team A
District Level "entry",
organization of 6 KIIs and
Team B
community entry and 4
household surveys in 1st
community of first cluster
Morning
KIIs: DDAs and BAC
afternoon
FGD: Generic Change Analysis,
men
branches)
(GPCs and SMEs)
FGD: Generic Change
6 household surveys in 1st
community of cluster 1
late
afternoon
Day 2 First Cluster
Day 3 Second
Cluster
Day 4 Second
Cluster
Day 5 "District
Level"
Evening
early
morning
household surveys in 2nd
Morning
women & youth
men
afternoon
participants in DSF /GPC
FFF
late
afternoon
third community entry and 5
household surveys in 3rd
community of 1st Cluster
third community entry and 5
household surveys in 3rd
Evening
early
morning
community of second cluster and
5 household surveys
community of second cluster
and 5 household surveys
Morning
men
FGD: Generic Change
afternoon
participants in DSF/GPC
FFF
late
afternoon
second community entry and 5
second community entry and 5
Evening
early
morning
women & youth
men
Morning
3rd community entry and 5
household surveys in second
cluster
3rd community entry and 5
household surveys in second
cluster
afternoon
late
afternoon
Evening
early
morning
Morning
Data entry
Data entry
172
afternoon
late
afternoon
Evening
Scenario 3: Nine days with only 3 clusters
Day
Location
Day 1 District level and
First Community
Day 2 First Cluster
Day 3 First Cluster
Day 4 Second Cluster
Time
early
morning
Team A
District Level "entry",
organization of 6 KIIs and
Team B
community
Morning
KIIs: DDAs and BAC
(GPCs and SMEs)
afternoon
(GPCs and SMEs)
community entry in 2nd
community of 1st cluster
late
afternoon
branches)
community entry in 3rd
Evening
early
morning
5 household surveys in 2nd
5 household surveys in 2nd
Morning
women & youth
men
afternoon
men
FGD: Generic Change
late
afternoon
5 household surveys in 3rd
5 household surveys in 3rd
Evening
early
morning
mopping up and preparation of
FGDs
mopping up and preparation
of FGDs
Morning
participants in DSF
FFF
afternoon
GPC
late
afternoon
Evening
early
morning
move to second cluster
move to second cluster
Morning
afternoon
late
afternoon
men
FGD: Generic Change
173
Day 5 Second Cluster
Evening
early
morning
women & youth
men
Morning
participants in DSF
FFF
afternoon
Day 6 Third Cluster
Day 7 Third Cluster
GPC
late
afternoon
Evening
early
morning
moving to next cluster
Data entry
community of third cluster and 5
household surveys
Data entry
community of third cluster
Morning
household surveys
afternoon
household surveys
late
afternoon
men
FGD: Generic Change
Evening
early
morning
women & youth
men
Morning
participants in DSF
FFF
afternoon
GPC
late
afternoon
Evening
early
morning
Morning
afternoon
late
afternoon
Evening
early
morning
Morning
afternoon
late
afternoon
Evening
Data entry
Data entry
Travel to district level (sensemaking workshop location)
Mopping up
Mopping up
174
Annex 16: Ethical principles and standards of conduct
The following is a framework of ethics observed by researchers that undertook data collection for this study.
It was discussed and agreed upon in the research training sessions. It is composed of 2 parts:


Part A: Research Ethics (standards which apply to any research at any time and in any location /
standard good research practice)
Part B: Community Ethics (standards which are specific to particular communities and cultural
contexts at a particular moment in time)
A. Research Ethics
1. Do no harm (physical, emotional, sexual) to anyone with whom you come into contact as a result
of this research, including research team members, respondents, community members, transport
providers and others.
2. Respect the rights of others:
 Research Team
 Respondents
 Community as a whole
3. Attitude: Be honest at all times and do not be afraid to admit if you do not know something; listen
actively; be polite and considerate; concentrate on the respondents; be punctual; recognise the
strengths of colleagues, respondents and community members; do your best to put respondents and
community members at ease.
4. Team work: Cooperate and be responsible to each other as a team; be flexible and help each other
out; conflicts of interest should be dealt with by referring them to the team leader (or other?) who
should raise the issue in a team discussion where appropriate in order to try to come to a consensus
of opinion; maintain good communication within the team.
5. Professional standards: Maintain professional boundaries – you are here to do research, not to
develop relationships with your co-researchers, respondents or community members; do not make
promises you cannot keep; be aware of what you can and can’t do; collect accurate data and check
your data before handing in completed tools to the team leader – you are responsible for the
information you collect and will be held accountable for errors.
6. Informed consent: Researchers must read out the informed consent instructions for each of the
research tools (adult and child household questionnaires, group activities and key informant
interviews) and sign the relevant informed consent sheets.
7. Confidentiality: This includes ensuring anonymity and confidentiality, in record keeping and
report writing and making sure participants understand that what they do and say in the group
session will remain anonymous. In addition, respondents should be made to feel at ease and
encouraged to equally ask researcher’s questions.
8. Impartiality: Do not show favouritism with colleagues, respondents and community members; do
not discriminate against anyone with whom you come into contact as a result of the research on the
basis of sex, religion, language, ethnicity, sexuality or any other grounds
9. Transparency: Decision-making and disciplinary processes must be transparent (i.e. everyone
should know how things are done and why decisions are made); the same process must be applied
to everyone, regardless of who they are; however, whilst processes are transparent, the information
that passes through these processes may be confidential.
10. Health and safety: The team leader, and ultimately the Research Manager, are responsible for the
health and safety of the research teams; research teams shall not be put at unnecessary risk; the
175
Research Manager and team leaders are expected to assess risks and to put in place steps to manage
risk; researchers should work in pairs; researchers should always know where other team members
are in case of emergency; do not move away from the team without telling others where you are
going and what time you will be back.
Below are some key ethical considerations to be made in carrying out participatory research:










Participant selection should be random and free of basis of, for example, access or stigma.
Permission is sought for the focus groups to go ahead, through consultation with the local
community.
Setting and communicating clear parameters for the focus group – this means clearly stating the
purpose, the limits and what the follow up will entail. It also means ensuring that demands on
participants’ time are not excessive and that they are aware of their right to not participate or
withdraw at any time.
Respondents were made aware of the fact that the research team are independent with no direct
associations with implementing agents
Setting up FGDs and interviews at a time and in places that are convenient to respondents (e.g.
after labouring hours)
Recognising that participants are possibly vulnerable and that the exercise is carried out with full
respect regardless of age or sex – power differentials will exist between community members and
researchers and these need to be purposefully mitigated in planning and implementation
Ensuring the safety and protection of participants – this means ensuring the environment is
physically safe, that there are at least two facilitators present at all times and, if possible, that a
local stakeholder group is involved in monitoring activities. Facilitators should also be supervised.
Ensuring that people understand what is happening at all time. Is appropriate language being used
(language, dialect, community terminology, etc)? This needs to be carefully planned.
Ensuring the right to privacy – this includes ensuring anonymity and confidentiality, in record
keeping and report writing and making sure participants understand that what they do and say in the
group session will remain anonymous. In addition, respondents should be made to feel at ease and
encouraged to equally ask researcher’s questions.
Not to harm ot threaten to harm any community member or colleague, regardless of age or sex,
either physically, sexually or emotionally.
B. Community Ethics
1. Dress code: Respect local norms and dress appropriately.
2. Hygiene: Maintain a good standard of personal hygiene.
3. Language: No swearing; use simple, polite language; no derogatory (bad or negative) words to be
used about people of another religion, ethnic group, province, sex, age etc.; avoid words which
have different meanings in different languages. This applies to the whole period of the research
both ‘on’ and ‘off-duty’.
4. Body language: No inappropriate or rude gestures; show respect and attentiveness in your body
language. This applies to the whole period of the research both ‘on’ and ‘off-duty’.
5. Physical environment: Consider the venue of the data collection to make sure it is safe, culturally
appropriate and as private as possible; care for the community environment and property – no
littering or vandalism.
6. Cultural context: Show respect for cultural practices, religious and community activities although if you witness something which is causing harm to a child then report this to the Team
176
Leader; all those involved with the research must observe culturally appropriate protocols during
preparation, entry into the community, during the research and exiting from the community.
C. Code of conduct
This Code of Conduct should be interpreted in a spirit of transparency and common sense, with the best
interests of the respondents as the primary consideration.
1. Research team members must make an attempt to understand the local norms around physical
contact.
2. Minimising risk situations:
 Try to: avoid placing yourself in a compromising or vulnerable position; be accompanied by a
second adult whenever possible; meet with a respondent in a central, public location whenever
possible; immediately note, in a designated organisational note book or incident report sheet,
the circumstances of any situation which occurs which may be subject to misinterpretation;
keep in mind that actions, no matter how well intended, are always subject to misinterpretation
by a third party.
 Try not to be alone with a single child, including in the following situations: in a car (no
matter how short the journey); overnight (no matter where the accommodation); in your home
or the home of a child.
3. Sexual behaviour:
 Do not: engage in or allow sexually provocative games with community members to take
place; kiss, hug, fondle, rub, or touch anyone in an inappropriate or culturally insensitive way;
use language that sexualises; create, view or distribute images in any format (print or
electronic) of anyone who is not appropriately clothed and / or who is depicted in any poses
that could be interpreted as sexually inappropriate
4. Physical environment:
 Do: develop clear rules to address specific physical safety issues relative to the local physical
environment of a project
5. Behaviour with community members and colleagues:
 Do: Treat all community members and colleagues, regardless of age or sex, with respect and
courtesy.
 Do not: Harm or threaten to harm any community member or colleague, regardless of age or
sex, either physically, sexually or emotionally. This includes - do not: hit (either with a hand or
other implement), intimidate, bully or sexually coerce or harass.
177
Annex 17: Approved Budget
The figures presented below were for 3 lead researchers and 9 assistants in 30 localities country-wide
(4 days per locality, thus a total of 40 days fieldwork).
A. FEES
DESIGN
Research Coordinator
Co-writer
Lead Researchers
Research assistants
Total
unit
DATA COLLECTION
Co-writer
Lead Researchers
Research assistants
Total
unit
ANALYSIS AND REPORT WRITING
Co-writer
Lead Researchers
Research assistants
Sub-Total
unit
22
5
15
8
cost/unit #
400
200
200
100
20
10
40
40
cost/unit #
400
200
200
100
25
10
15
10
cost/unit #
400
200
200
100
total
1
1
3
9
8,800.00
1,000.00
9,000.00
7,200.00
26,000.00
total
1
1
3
9
8,000.00
2,000.00
24,000.00
36,000.00
70,000.00
total
1
1
3
9
SUB-TOTAL:
10,000.00
2,000.00
9,000.00
9,000.00
30,000.00
$126,000.00
B. REIMBURSABLES
DESIGN
Transportation to Kumasi
Accommodation lead researchers + coordinator
Accommodation co-writer + research assistants
Per Diem lead researchers + coordinator
Per Diem co-writer + research assistants
Consultation WS -transport participants
Consultation WS -accommodation participants
Consultation WS -conference package participants
Stationary, materials, printing & copies
Car fuel
Sub-Total
unit
1
7
0
14
7
1
1
1
1
4
cost/unit #
$35.00
$50.00
$50.00
$20.00
$20.00
$50.00
$50.00
$35.00
$100.00
$30.00
total
4
4
10
4
9
8
8
30
1
4
$140.00
$1,400.00
$0.00
$1,120.00
$1,260.00
$400.00
$400.00
$1,050.00
$100.00
$480.00
$6,350.00
178
DATA COLLECTION & PROCESSING
Per Diem research coordinator
Per Diem co-writer
Per Diem lead researchers + assistants + drivers
Accomodation research coordinator
Accomodation co-writer
Accommodation lead researchers + assistants +
drivers
Interpreters
Community entry
Community sensemaking workshops
Car fuel
Quant + Qual Data entry
Sub-Total
unit
ANALYSIS & REPORTING
Transportation to Kumasi
Per Diem research team
Accommodation research team
Consultation WS -transport participants
Consultation WS -accommodation participants
Consultation WS -conf package participants
Sub-Total
unit
SUB-TOTAL:
20
0
40
20
0
40
cost/unit #
$20.00
$20.00
$20.00
$40.00
$40.00
$40.00
1
1
15
1
1
15
$400.00
$0.00
$12,000.00
$800.00
$0.00
$24,000.00
40
1
40
1
40
1
$20.00
$20.00
$2.50
$500.00
$30.00
$2,000.00
6
30
30
1
3.5
1
$4,800.00
$600.00
$3,000.00
$500.00
$4,200.00
$2,000.00
$52,300.00
1
5
5
1
2
3
1
cost/unit #
$35.00
$20.00
$50.00
$35.00
$50.00
$35.00
$500.00
total
total
14
14
14
130
130
130
1
$490.00
$1,400.00
$3,500.00
$4,550.00
$13,000.00
$13,650.00
$500.00
$37,090.00
$95,740.00
C. TOTAL:
Fees
Reimbursable
SUB-TOTAL:
Overhead 5%
$126,000.00
$95,740.00
$221,740.00
$11,087.00
TOTAL:
$232,827.00
179

Final report on the participatory impact evaluation of the Root

Transcription

Similar documents

Russell

Cassava Hornworm Outbreak

Agriculture in Romania

your promise to the world

January 2014

New Food Rules - The Health Trust

sale october 15 thru october 28, 2012

Shelby County - Alabama Farmers Federation