Joint Publication 9 
CARE Madagascar 
Title II Final Evaluation 
Report of Findings 
Prepared by Richard Caldwell, 
TANGO International, Inc. 
and Sarah Fry, 
USAID-funded Environmental Health Project 
January 2003 
CARE Madagascar: Title II DAP Final Evaluation Page i 
Table of Contents 
Acronyms.......................................................................................................................... iii 
Executive Summary......................................................................................................... iv 
I. Introduction......................................................................................... 1 
II. Objectives of the Evaluation.............................................................. 2 
III. Evaluation Methodology .................................................................... 5 
 III.A Review and Analysis of Key Documents...............................................5
 III.B Key Informant Interviews......................................................................6 
 III.C Primary Data Collection in the Field ....................................................6 
 III.D Project Site Selection and Participant Interviews ...............................7 
 III.E CARE Staff and Counterpart Participation in the 
 Evaluation Process..................................................................................7 
IV. Evaluation Findings............................................................................ 8 
 IV.A Mahavita..................................................................................................8 
 IV.A.1 Strategic Objective 1 .............................................................................8 
 Income Generation and Assets 
 Training – Personal Empowerment and Livelihoods 
 Savings Component 
 FFW Activities 
 IV.A.2 Strategic Objective 2 ...........................................................................14 
 Developing Community-based Structures 
 Fokotany Selection
 Empowering Structures de Developpement 
 SD Activities and Management 
IV.A.3 Strategic Objective 3 ...........................................................................18
 Infrastructure Development 
 Water/Sanitation/Washbasins 
 Hygiene Promotion 
 Enabling Environment 
 IV.A.4 Strategic Objective 4 ...........................................................................23 
 Health Sector 
 Malnutrition 
 IV.B CYPREP (Cyclone Preparation on the East Coast) ..........................25 
 IV.B.1 Intermediate Result 1 ..........................................................................25 
 Strategic Orientation 1:
 Strategic Orientation 2: 
 Fokontany Development Plans (PDF) 
 Cyclone Preparedness Plans (PPCs) 
 Strategic Orientation 3: 
CARE Madagascar: Title II DAP Final Evaluation Page ii 
 Mitigation Activities 
IV.B.2 Intermediate Result 2 ......................................................................31 
 Strategic Orientation 4: 
 Strategic Orientation 5: 
 Accomplishment of CVS Tasks 
 IV.B.3 Intermediate Result 3 ......................................................................33 
 Strategic Orientation 6: 
 The Commune Emergency Management Committee 
Capacity-building of Social Mobilization Agents 
 IV.C Food Aid Management Systems & Internal Controls .......................35 
 IV.D Food Distributions in FFW Programs ................................................35 
V. Conclusions................................................................................... 36 
V.A Mahavita................................................................................................36 
 Strategic Objective 1: 
 Strategic Objective 2: 
 Strategic Objective 3: 
 Strategic Objective 4: 
 V.B East Coast Programming – CYPREP.................................................39 
 Intermediate Result 1: 
 Intermediate Result 2: 
 Intermediate Result 3: 
 V.B.1 Other CYPREP Conclusions ...............................................................43 
VI. Recommendations for DAP II .................................................... 43 
 Urban Development/Governance 
 Environmental Health 
 Health 
 Guiding Principles for DAP II Environmental Health activities 
 Enhancing Food and Livelihood Security on the East Coast 
Annexes 
Annex A CDD (James Garrett) ....................................................................................50 
Annex B Evaluation Terms of Reference ....................................................................59 
Annex C List of Persons Contacted .............................................................................65 
Annex D Mahavita Fokotanys......................................................................................66 
Annex E Qualitative Survey Sites for CYPREP..........................................................67 
Annex F Qualitative Research Site Visits ...................................................................68 
Annex G Engineering Report (in French) ....................................................................70 
CARE Madagascar: Title II DAP Final Evaluation Page iii 
Glossary and Acronyms 
ADC Agent de Développement Communautaire 
AMS Assistant de Mobilisation Sociale (Field/Extension Agent) 
APIPA Autorité pour la Protection contre les Inondations de la Plaine 
d’Antananarivo 
CARE Cooperative for Assistance and Relief Everywhere 
CCS Conseil Communal de Secours (Commune Emergency Management 
Committee)
CFW cash-for-work 
CLS Conseil Local de Secours (niveau sous-préfecture) (District Emergency 
Management Committee)
COGES Comite de la Gestion et Suive (Management and Monitoring 
 Committee) 
CYPREP Cyclone Preparation Project 
CUA Commune Urbaine de Antananarivo 
CVS Comite Villageois de Securite (Village Security Committee)
DAP Development Assistance Program 
EHP Environmental Health Project (USAID) 
FALAFA Famakafakana Lalina ny Fiainana eny Ambanivohitra (Malagasy acronym 
for HLS analysis)
FAMOA Fanadihadiana Miaraka ny Olana sy ny Alaolana 
FFP Office of Food for Peace
FKT Fokotany (the lowest administrative unit in the Government of 
 Madagascar) 
FMG Francs Malagasy (approximately 6,200 FMG = $1.00 U.S.) 
Fokontany Malagasy word for village or community
FFW food-for-work 
GOM Government of Madagascar 
HIF Hygiene Improvement Framework 
HLS Household Livelihood Security 
IFPRI International Food Policy Research Institute 
ISIKI Io no Sehatra Ifanakalozan-Kevitra sy Andrindran’Asa 
LOP life of project 
MOH Ministry of Health 
PAIQ Programme d’Appui aux Initiatives de Quartier 
PDF Plan de Developpement du Fokotany (Fokotany Development Plan) 
PDQ Plan de Developpement du Quartier (Neighborhood Development Plan) 
PL Public Law (480) 
PPCC Plan de Préparation Cyclonique Communal (Commune Cyclone 
Preparedness Plan)
SCVM Sécurité des Conditions de Vie des Ménages (HLS)
SD Structures de Developpement 
USAID United States Agency for International Development 
VCT vivre contre travail (food-for-work) 
CARE Madagascar: Title II DAP Final Evaluation Page iv 
Executive Summary 
CARE Madagascar’s Title II DAP began in 1998 and included two sub-programs - Mahavita and 
CYPREP. Mahavita was operational in 30 of Antananarivo’s poorest neighborhoods and aimed 
at sustainably improving household food and livelihood security. Its primary delivery 
mechanism was a well-orchestrated, community-based approach. CYPREP assisted rural 
communities throughout Tamatave Province on the east coast of Madagascar in risk management 
through planning and preparing for natural disasters. This evaluation provides findings from the 
evaluation of both components, as well as conclusions regarding the impact of the program and 
recommendations for future work in both rural and urban settings. 
The program set out to facilitate the creation of community-based structures (SDs) empowered to 
identify, plan, implement, coordinate and sustain activities that would contribute to the well￾being of residents. In this Mahavita made significant progress in empowering communities and 
the majority completed community plans to guide development over the coming years. Many 
residents express a feeling of being part of an important process and that their efforts will result 
in improved living conditions. 
Mahavita’s efforts to improve income and savings in poor households could have been done 
more effectively, and as a result there is little evidence that incomes have increased due to new 
employment opportunities for the urban poor. The FFW component, however, was very 
successful in targeting poor households and in providing a safety net in the form of cash and 
food. This was particularly critical during the political crisis of 2002 when food wages made an 
important economic contribution to households. 
Most of the community SDs are now struggling with ‘next steps.’ Members still appear highly 
motivated, and some SDs are now seeking their own funding with limited success, but there has 
been no coordinated effort on the part of Mahavita to develop alternative strategies and capacity 
for such action. This will be a key area to reinforce in the follow on DAP. 
Mahavita also facilitated improvements in the environmental health of Fokotanys 
(neighborhoods) through training and infrastructure development. Infrastructure development 
followed community empowerment, and it is more likely infrastructure will be maintained by the 
community. Where program activities have been active the neighborhoods are noticeably 
cleaner. Residents speak of noticeable health improvements in areas such as skin disease, 
children’s diarrhea, and even infant mortality. Despite positive gains in environmental health, 
more work is needed in water and sanitation, especially as it applies to fecal matter disposal. 
The CYPREP component was successful in developing networks of local partners for disaster 
management and response and for mobilizing organizations and communities around important 
livelihood themes. Communities are now organized to warn residents of impending storms and 
plan for how they can mitigate the impact of disasters. Communities also for the first time have 
development plans that not only address disaster response, but also incorporate improvements to 
livelihoods. In addition, there is much more understanding by community members concerning 
the impact of cyclones and the action needed to mitigate them.
CARE Madagascar: Title II DAP Final Evaluation Page 1 
FINAL EVALUATION 
CARE MADAGASCAR TITLE II DEVELOPMENT ACTIVITY 
1998-2003 
I. Introduction 
Households in Madagascar’s urban neighborhoods face critical food and livelihood security 
problems experienced in many other of the world’s poorest urban environments. Among the 
challenges household’s face include a changing and often unstable economic climate, rapid 
population growth, inefficient and poorly functioning municipal services, and intense 
competition for limited income-generating opportunities. The urban poor have aspirations like 
any other household – they want to send their children to school, live in a comfortable home, eat 
preferred foods, and maintain good health. All of these aspirations present challenges to the 
majority of Antananarivo’s urban poor as they try to achieve stable food and livelihood security. 
CARE Madagascar uses U.S. Government resources to address food security problems in 
Antananarivo, Madagascar’s capital. Its current Title II DAP was submitted, off-cycle, in 
September 1997 and includes two sub-programs: Mahavita1
 and CYPREP. Mahavita began in 
1998 and is an urban household food and livelihood security program. It assists the populations 
of 30 Fokontany (FKT), the lowest formal administrative structure in the Government of 
Madagascar. Each FKT represents 8-10,000 people and is located within Antananarivo’s urban 
flood zone, an area subject to complete inundation of water for 3-4 months per year and 
representing the most insalubrious environments in the city. The total population of these thirty 
FKTs is approximately 200,000 and is comprised mainly poor urban households, some of which 
are squatters, recent migrants from rural areas, and economically displaced urbanites. 
Mahavita has evolved as an umbrella program with an overall goal of sustainably improving 
household food and livelihood security of poor households in Antananarivo’s most vulnerable 
communities. It is comprised of interventions designed to improve household income and 
savings, community and personal empowerment, infrastructure and health and hygiene behavior. 
Its primary delivery mechanism is a well-orchestrated, community-based approach. Projects 
included under Mahavita include PAIQ (Dec. 1995 – Sept. 1997)2
, TOUCH 2000 (Feb. 1996 to 
present) and Safe Water System. The main objectives of these projects are included in Table 1. 
CARE Madagascar’s Title II program’s CYPREP component assists rural communities 
throughout Tamatave Province on the east coast of Madagascar. This cyclone-prone coastal 
region is rich in natural resources yet poor households are highly vulnerable to severe storms 
each year. CYPREP assists communities in risk management through planning and preparing for 
natural disasters. Through efforts over the past three years, CARE has established strong 
 
1
 Mahavita is a Malagasy term meaning “working together to achieve a common goal.” 2
 Though this project ended before Mahavita began, PAIQ project efforts were retained and continued to be a large 
part of Mahavita. 
CARE Madagascar: Title II DAP Final Evaluation Page 2 
relationships with fokotany, district, and provincial governments in its cyclone awareness 
program. 
Table 1. Sub-components included in Mahavita. 
Projects Dates Objectives 
PAIQ Dec. 1995-Sept. 1997 • Provide technical assistance to local communities in 
Antananarivo for the submission, funding, implementation, 
follow-up and post-project management of small urban 
infrastructure projects. 
TOUCH 2000 Jan. 1996-September 
1999 
• Reduce infant and child mortality in poor urban areas of 
Antananarivo through child survival interventions at public and 
private health facilities and at community and household levels. 
Safe Water 
System 
Oct. 1999 – Aug 2001 • Improve household water quality and decrease diarrheal disease 
in the target population by promoting treatment of contaminated 
water, safe water storage in plastic containers, and behavior 
change techniques, including social marketing, communication 
and education. 
Antananarivo 
Emergency 
Program 
January-July, 2002 • Provided a variety of FFW opportunities for households during 
the political crisis of 2001 
AGETIPA Ongoing • Co-financed infra-structure in Mahavita fokotanys 
Petit Boulevarde July 2003 - present • Mahavita provides assistance on community awareness and 
social conflict resolution and supports urban water 
infrastructure activities. 
This report will discuss results from both the Mahavita and CYPREP components. It is largely 
organized around the strategic objectives of each component and includes observations, findings, 
and ways forward for future urban and East coast programming. 
II. Objectives of the Evaluation 
The DAP’s Monitoring and Evaluation System stipulated that: 
CARE-Madagascar’s DAP activities will be assessed through mid-term and final evaluations, 
which will cover both project effectiveness and impact. The final evaluation will be 
conducted in the final year of the project. Unlike the participatory approach used for the 
mid-term evaluation, the final evaluation will be conducted by an outside agency in 
compliance with PL-480 Title II DAP guidelines. 
The primary objective of the evaluation was to determine the effectiveness of the approaches 
used by Mahavita and CYPREP as well as the effect3
 and impact changes that were brought 
about as the result of activities and outputs achieved throughout the life of project. The 
evaluation included several phases and components and was conducted by a team of three 
 
3
 Here “effect” means the behavioral and systemic changes brought about as a result of achieving project-related 
outputs. 
CARE Madagascar: Title II DAP Final Evaluation Page 3 
consultants, two international and one national4
. It has also greatly benefited from two other 
pieces of work. The first was a case study of community development by James Garrett of the 
International Food Policy Research Institute (IFPRI) entitled “Scaling up Community-Driven 
Development: Reflections on the Mahavita Experience.” Since this study highlighted many of 
the processes used by Mahavita it is being used as a key supporting document in this evaluation 
and is included as Annex A. 
The second contribution was a commentary by Michaell Drinkwater of CARE on how Mahavita 
can transition into its next phase of urban programming and is entitled “Building Urban 
Governance in Antananarivo: Looking to the Future of the Mahavita Program” (CARE 2003). 
This document is available from CARE Madagascar. 
For Mahavita, the analytical focus of the evaluation was to assess the following questions: 
y To what degree have the original goals, objectives, and project outputs been achieved, 
and what are the main factors that contributed to the level of achievement? 
y How appropriate was the overall design of the program and of the specific objectives, 
activities and strategy to address the root causes of vulnerability and food insecurity? In 
addition, how realistic were the life-of-project and annual targets, the staffing plan and 
strategy, information systems (including monitoring and evaluation systems as well as 
geographic information systems), and budgets?. 
y How effective was program implementation, including the strategy and methodology 
employed, staffing effectiveness, and the quality of coordination and collaboration with 
project counterparts. 
y What was the quality and sustainability of specific interventions, including the project’s 
success in accomplishing planned results? What was the effectiveness of community 
mobilization and empowerment, and the effectiveness of Structures de Developpements
(SD) in planning and coordinating community-driven activities? 
y Were there any unintended impacts, including applied mitigation measures, and internal 
and external factors that affected program implementation and outcomes, either in a 
positive or a negative manner? 
The environmental health component determined whether the program had been able to attain its 
stated objectives of improving environmental health and hygiene behaviors at the community 
and the household levels. A staff member from USAID’s Environmental Health Project (EHP) 
was invited to participate in the evaluation and proposed to conduct its part according to the 
Hygiene Improvement Framework (HIF) model, a program organizing principle for diarrhea 
prevention advocated by EHP, UNICEF and others. The HIF has three components that need to 
be addressed together in order to achieve the desired result of disease prevention: 
 
4 The consultants for the Mahavita components were Richard Caldwell (Team Leader, TANGO International, Inc) 
Sarah Fry (USAID Environmental Health Project), A. Andriamasy (Infrastructure). For CYPREP the main 
consultants were Rakotomalala Bonaventure (ESA, qualitative research) and Andrianaivo Heritiana (quantitative 
research). 
CARE Madagascar: Title II DAP Final Evaluation Page 4 
y Hardware (including water/sanitation infrastructure improvement and access to 
household technologies such as soap, water treatment materials, children’s potties) 
y Hygiene promotion (IEC, social mobilization, community participation, advocacy) 
y Strengthening the enabling environment (policy improvement, institutional strengthening, 
community organization, financing and cost-recovery, cross-sector and public/private 
partnerships). 
The evaluation mostly focused on the first two components (hardware/technologies and hygiene 
promotion) since a review of the enabling environment was a key part of the team leader’s Scope 
of Work. However, the enabling environment as it relates directly to environmental health was 
considered, for example, community management and cost-recovery systems established for 
infrastructure maintenance. Assessment of the quality and sustainability of the infrastructure 
itself is part of the consultant engineer’s SOW. 
During the course of the evaluation, the importance of CARE’s original involvement with the 
Ministry of Health (MOH) became apparent, and some time was devoted to exploring the history 
and status of relations as well as possible future connections, with MOH representatives at 
various levels. 
The infrastructure component involved cash and later food for work to rehabilitate canals and 
construct walkways, and provide water provisioning through cost-recovery systems. The 
evaluation examined these infrastructures by tabulating the outputs (# person-days, kilometers of 
canals rehabilitated, etc.), conducting physical inspections, and by reviewing processes and 
procedures, both social and technical. 
The fourth component involves evaluating the outputs of CYPREP and the effectiveness of the 
methodology used. In addition, CARE would like to use this evaluation as a means of preparing 
for the start-up of the new DAP. 
For CYPREP, the analytical focus of the evaluation was to assess: 
y Accomplishment of the goals, objectives, and project outputs. 
y Program design, including the appropriateness of the specific objectives, activities and 
strategy to address the root causes of vulnerability. In addition, annual targets, staffing 
plan, information systems, and budgets were addressed. 
y Program implementation, including the strategy and methodology employed, staffing 
effectiveness, and the quality of coordination and collaboration with project counterparts. 
y The quality and sustainability of specific interventions, including project success in 
accomplishing the results, the effectiveness of community mobilization and 
empowerment, and the effectiveness of community development structures in planning 
and coordinating community-driven activities. 
y Unintended impacts, including applied mitigation measures, and internal and external 
factors that affected program implementation and outcomes.
CARE Madagascar: Title II DAP Final Evaluation Page 5 
The terms of reference for the CYPREP and Mahavita final evaluation are included as Annex B. 
III. Evaluation Methodology 
A variety of activities were included in the evaluation to make it as participatory as possible and 
to broaden the range of viewpoints to be included. This included group meetings with Mahavita 
staff., interviews with CARE staff (ADC, IEC, Supervision, Infrastructure, Income Generation), 
informal talks with CARE and Mahavita management (present and past). A list of persons 
contacted is provided in Annex C. Evaluation team members held many meetings with health 
and municipal officials, including representatives at the communal level and the Arrondissement 
level. There were numerous field visits made to Fokotany’s where group discussions were held 
with Structures de Developpement (SDs) and FKT representatives. Other less formal 
discussions were held with staff and residents during neighborhood observational walks and 
physical “inspection” (technical and non-technical) of infrastructure. 
III.A. Review and Analysis of Key Documents
The evaluation team reviewed several key CARE-Madagascar documents, including: 
y The original CARE Madagascar DAP, submitted in 1998; 
y Annual results reports from FY99 – FY2002; 
y Mahavita ‘Start-up Process Document’; 
y CARE-Ethiopia’s Long Range Strategic Plan of July 2001 to June 2006; 
y Mahavita Baseline Survey Document, 2000; 
y Mahavita Final Survey Document, 2003; 
y FAMOA – community mobilization and institutional capacity building 
strategy/methodology document; 
y HIMO – food for work planning, implementation and monitoring document; 
y ISIKA – Coordination and capacity building of stakeholders in water and sanitation 
infrastructure; 
y Mahavita Annual Evaluation and Planning Exercises; and 
y FITARIHA – Monitoring System 
y Food Aid Management Review 
y CYPREP Annual Reports 2001 and 2002
y CYPREP 2002 Work Plan
The evaluation team also reviewed several specific project documents in the Mahavita and 
CYPREP field offices. These included miscellaneous project documents such as trip reports as 
well as proposals and progress reports for related activities (e.g. - TOUCH 2000 and Sur Eau). 
III.B. Key Informant Interviews 
In order to accumulate multiple perspectives, team members conducted key informant interviews 
with a wide variety of stakeholders involved in the DAP, including: 
CARE Madagascar: Title II DAP Final Evaluation Page 6 
y CARE Madagascar program staff and management; 
y USAID Food-for-Peace Officers; 
y Fokotany officials; 
y Community members; 
y Arrondissement and Commune officials; 
y CYPREP Regional Managers in Tamatave; and 
y CYPREP Field Managers in Mahanoro. 
III.C. Primary Data Collection in the Field 
The most substantial and essential data for the final evaluation was collected during project visits 
to the field. For Mahavita, these visits were targeted to urban residents, members of SDs, and 
Fokotany leaders. It was not possible to visit each of the thirty Fokotany’s, but time did allow a 
significant number and range of sites to be visited, and included areas where both social and 
physical changes were being made and where current food-for-work activities are taking place. 
Annex D provides a list of the thirty Fokotanys included in Mahavita and denotes those that were 
visited. 
For CYPREP, field discussions were held with project engineers, Communal leaders, Fokotany 
leaders, members of CVS, and community members. Quantitative data was collected from 
household members. Annex E provides a list of sample sites from the qualitative survey and 
Annex F provides details of qualitative research site visits. 
CARE Madagascar: Title II DAP Final Evaluation Page 7 
III.D. Project Site Selection and Participant Interviews 
In the urban context, a variety of Fokotanys were selected for inclusion in the evaluation study 
on the basis of: 
y Age & history of CARE involvement in the Fokotany; 
y Socio-economic status of the Fokotany (relative poverty or affluence as evidenced 
by the 2003 final survey); 
y Types of activities implemented within the Fokotany; and 
In the rural context of CYPREP, a variety of Fokotanys were selected for inclusion in the 
evaluation study on the basis of: 
y Age & history of CARE involvement in the Fokotany; 
y Level of vulnerability and ability to match with non-CYPREP villages; 
y Types of activities implemented within the Fokotany; and 
Each field visit consisted of a combination of: 
y Key informant interviews with Fokotany administration officials; 
y Focus group interviews with the SD – the development committee responsible for 
community-driven development; 
y Focus group interviews with foor-for-work participants; 
y Interviews with Fokotany residents; and 
y A tour of the Fokotany project activity site to assess technical quality and 
sustainability of specific infrastructure interventions. 
III.E. CARE Staff and Counterpart Participation in the Evaluation Process
The evaluation team spent considerable time interviewing CARE project management and staff, 
including field agents. Additional time was devoted to interviewing … 
Finally, the evaluation team spent half a day with CARE staff to present the results of the data 
collection, discuss the accuracy and relevance of the findings, and together analyzing the 
findings. These were valuable exercises that brought closure to the evaluation team’s visits to 
urban and rural sites and added new perspectives to the findings. 
CARE Madagascar: Title II DAP Final Evaluation Page 8 
IV. Evaluation Findings 
IV.A Mahavita 
Program Mahavita’s is somewhat unique among CARE’s projects in Southern and Western 
Africa and among Title II DAPs in general. During the first year of operation it applied a 
number of theoretical concepts and focused on building organizational structure and establishing 
common vision and principles. This “inception phase” was a well-crafted strategy to accomplish 
solid team building, empower staff, and understand communities and the operational 
environment of Antananarivo well enough so that effective programming could be achieved. 
The second year, FY 2000, Program Mahavita activities were mainly geared towards designing 
and testing methodologies for program implementation through a ‘pilot phase’, which also 
served as a practical training experience – ‘learning-while-designing’ – for the program team. 
The ‘pilot phase’ proved to be an effective means for achieving quality outputs in a relatively 
short period of time. It fostered group learning, and emphasized the importance of teamwork. It 
was also an effective way for the program team to gain ownership of processes. 
While it did not achieve all it set out to do, Mahavita was a pivotal project in the evolution of 
Title II programming. It was perhaps a few years ahead of its time in terms of focusing on issues 
such as empowerment and community mobilization in an urban food security context5
 and would 
fit quite comfortably with the new Office of Food for Peace strategy that was recently 
articulated. Mahavita operated around three strategic objectives related to improving the food 
security of poor urban households in Antananarivo. The first addressed food access by focusing 
on income and savings. The second objective was to empower communities to address 
constraints affecting their ability to control their own resources and livelihoods. The third 
objective was to create and maintain healthy and hygienic environments that could support 
productive livelihoods. The fourth and final objective was to promote healthy behaviors at the 
household level. The findings within each of these objectives are discussed below. 
IV.A.1 Strategic Objective 1 
Targeted households will be using increased income and savings to provide for their 
food and livelihood needs. 
Income Generation and Assets
Income generation was recognized in the original project document as important to urban 
livelihoods. Indeed, income was the focal point of problem analyses conducted during the 
design of Mahavita and was one of the four objectives of the original TOUCH 2000 project. 
Improving household incomes was to come from several interventions, including food-for-work 
 
5
 Mahavita, although designed five years ago, is fully compatible with the new strategy document put out by the 
Office of Food for Peace, which has a number of Intermediate Results based on community capacity building, 
resilience, and human capacity building. 
CARE Madagascar: Title II DAP Final Evaluation Page 9 
activities as a temporary income transfer and skills training that would enable individuals to 
better compete in a tight employment market. It was also recognized that in addition to training 
and education, individuals had to develop the confidence and capacities that would allow them to 
take control over their own livelihoods and manage interrelationships with others. Thus the 
personal empowerment elements of the project became integral to the strategy of income 
generation. 
Food-for-work was part of the overall income generating strategy of Mahavita. Originally, those 
recruited to work on infrastructure activities would be given 12-18 month contracts and paid in 
cash or food. Food received as wages would partly substitute for income during the 
infrastructure work and would improve household food consumption. During their time in the 
CFW or FFW activities, participants would also be involved in training activities, particularly 
during the rainy season months when infrastructure work is halted. These training activities 
were designed to help participants initiate and manage income generating activities more 
successfully. Thus, a second income generating strategy of Mahavita was to graduate 
community members out of skills and life training and encourage them to seek their own 
opportunities. Improving incomes was to include micro-enterprise development as well as skills 
training for improving individual employment prospects in the formal sector. It was envisioned 
that part of the population would take advantage of these opportunities and overall incomes 
would improve. 
It is unfortunate to address income generation as the first finding of the evaluation, because this 
was perhaps Mahavita’s major shortcoming in an otherwise successful project. The only focused 
income-generating strategy that reached a significant percentage of the targeted population was 
food-for-work. While there is no doubt that participation in FFW relieved households of some of 
their economic burden and substituted partly for income needs, and thus had a positive outcome 
on food security, it is questionable whether it had a tangible impact on incomes in the long run 
for the vast majority of households. It is also debatable whether or not it could have a 
sustainable positive impact on a household’s ability to access food since in an urban context 
income is so key to addressing access issues. 
There is some evidence, however, to suggest that the FFW experience could be effectively 
parlayed into viable skills training, and that some of the more entrepreneurial members may be 
able to generate work by offering their labor to service providers. There was one case of 
employment generated by SD members with skills acquired by FFW and applied in a private 
sector, contractual approach. The first occurrence was in FY2001 when 8,000 person days of 
employment was generated. For that particular year this 200 percent more than the target of 
3,750 work-days. However, the LOP goal was 33,750 person-days. Thus, the overall 
achievement in this area was only 24 percent of that planned. The work experience that many 
gained was a positive and long-term effect, but to generate sustained employment requires a 
more intensive strategy than was used in Mahavita. 
In order to measure changes in the economic status of households asset ownership was selected 
as a proxy indicator, as previous studies in Madagascar and elsewhere had demonstrated the 
linkage between asset ownership and household wealth. Since direct income measurements were 
deemed unreliable, the asset status of the household was used to assess economic changes. A 
CARE Madagascar: Title II DAP Final Evaluation Page 10 
standard set of assets was measured for each household. In the baseline survey these were 
measured for the general population from a random sample of households. In the final survey, 
these same measurements were made from samples of FFW households, households from 
Mahavita Fokotanys, and households from non-Mahavita Fokotanys. The results are provided in 
Table 2. 
Asset ownership in the baseline survey averaged about 2,200,000 FMG per household. In the 
final survey, the mean asset ownership for FFW households averaged significantly less at about 
1,123,000 FMG. For non-FFW households the value rose about 30 percent to almost 3,000,000 
FMG. Comparison household assets valued nearly 3,500,000 FMG. FFW targeted the poorest 
households, so baseline figures are not directly comparable with final survey figures. It does 
confirm, however, the FFW households were by and large poorer. It is most likely that asset 
ownership has not significantly changed for the overall population. Inflation has hovered at 8-10 
percent per annum6
 during the life of the project, and when factored in to asset value the 
differences between non-FFW households and baseline households are not significant. 
There is also no evidence that FFW households are better off than similar households in 2000. In 
2000, those households in the same asset category as FFW households averaged about 950,000 
FMG in assets. In the final survey they averaged just over 1.1 million FMG in assets. Again, 
when inflation is factored in, these two values converge, so that FFW households probably have 
about the same value of assets they did in the beginning of the project. 
Table 2: Asset ownership, in FMG, by household type. 
HH Type Electronics Furniture Utensils Appliances Vehicles Total 
Assets 
mean 255,990 523,791 308,519 17,626 17,208 1,123,135 FFW s.d. 703,430 1,180,708 1,227,404 60,316 116,666 2,847,151 
mean 807,768 1,296,875 431,642 217,621 199,396 2,953,507 Non-FFW s.d. 1,366,443 2,185,848 885,804 759,102 982,000 4,389,986 
mean 936,258 1,471,482 281,224 236,541 573,405 3,498,912 Comparison s.d. 1,538,831 2,938,621 561,017 725,714 3,391,136 6,438,718 
mean 639,366 629,691 144,145 161,003 649,400 2,223,607 Baseline s.d. 1,551,798 1,046,314 211,835 643,138 3,039,979 5,197,099 
This being said, there was a significant and prolonged political crisis during the latter stages of 
the project, and it is almost certain that FFW acted as an important safety net for FFW 
households. Employment opportunities during the six month crisis were even fewer than normal 
while at the same time food prices rapidly rose as a consequence of transportation problems 
(mainly a lack of fuel) and the blockage of goods from the port in Tamatave. During this time 
FFW was one of the few alternatives, if not the only one, that most poor households had for 
employment and for accessing food. The FFW activities provided an important safety net for 
 
6 Azam, Jean-Paul. 2000. Inflation and Macroeconomic Instability in Madagascar. ARQADE and IDEI, University of 
Toulouse, and Institut Universitaire de France. June 2000. 
CARE Madagascar: Title II DAP Final Evaluation Page 11 
Mahavita’s poorest households, and it may have resulted in significant asset retention for these 
households or prevented them from having to take out expensive consumption loans to meet their 
food needs. Qualitative evidence supports this hypothesis, as many of the FFW participants 
relate that the employment FFW generated was the only thing that ‘saved’ them during this 
stressful time period. Asset retention in this context was as important as asset accumulation, for 
if poor households had been forced to sell their meager assets to cope with economic shocks they 
would have become more vulnerable, and the recovery period of the poorest households could 
have been quite long. 
Training – Personal Empowerment and Livelihoods
Training was a major activity of Mahavita and an important intervention for poor households. 
The opportunities to receive formal or non-formal training for adults of poor households are 
almost non-existent, save for projects like Mahavita. A total of 16,892 individuals benefited 
from personal empowerment training, which represents 169 percent of Mahavita’s original target 
and a significant percentage of the adult population in the thirty Fokotanys where Mahavita 
worked. 
Personal empowerment training and associated support activities were based on earlier CARE 
work in Zambia, and were developed to encourage the formation of mutual support and interest 
groups, and to assist these groups in identifying key livelihoods constraints of an economic and 
social nature (including gender issues) that inhibited households from engaging in productive 
activities. The types of activities envisioned included the formation of savings groups and 
targeted income-generation. 
Mahavita trainers worked with groups to elicit ideas and strategies for personal improvement, 
including skills training. Personal empowerment followed, focusing on enhancing people’s 
confidence and improving their interpersonal, business and decision-making skills to initiate 
activities. This included working with groups to commence savings activities, to improve 
existing income generating activities, or to identify and start-up more profitable new activities. 
It was found that for the poorest community members, improving their democratic voice 
addressed the issue of their exclusion from decision making processes, but in and of itself was 
insufficient to impact their employment prospects and other livelihood concerns. Before true 
gains could be achieved, the project had to build human capital for the poorest groups. 
Approximately forty percent of project participants have been able to upgrade their skills. For 
females, the most common income-related training was in canal and dike maintenance and 
rehabilitation. The training itself was in conjunction with FFW activities. A high percentage of 
FFW participants have never attended school (49 percent of FFW household heads, for example, 
have never attended school, compared with 20% of non FFW household heads and 27% of 
comparison households). Project staff as well as ex-FFW laborers commented that residents 
have taken greater initiatives to search for jobs within and outside the communities. Project 
personnel believe that personal empowerment training could be one of the most substantial 
sustainable effects of the project. However, as noted above, skills improvement has not 
necessarily transferred to income/asset enhancement due to a lack of employment opportunities 
CARE Madagascar: Title II DAP Final Evaluation Page 12 
outside of FFW. It has, though, contributed to the personal empowerment of individuals, 
especially females, and provided new avenues of opportunities that perhaps can be pursued over 
time. 
Table 3 shows the types of skills training received by community members during the last five 
years that participated in the final survey. Obviously not all of these people were trained by 
Mahavita, but it is evident that for the FFW households the opportunities provided by Mahavita 
were significant (for example, almost 60 percent had received training in various aspects of 
infrastructure maintenance). These are, however, probably not the skill sets that will result in 
significant employment opportunities. 
A majority of laborers and training recipients have been women, enhancing short-term household 
food security by putting food commodities under the control of women. Women have also been 
employed in many of the management positions as foremen, timekeepers and commodity issuers, 
but few have been used in skilled labor positions such masonry. 
Savings Component
The DAP design recognized household savings as key to achieving sustainable improvements in 
income and consumption, along with training and gender facilitation activities, and the 
strengthening of community based organizations which address rights and access issues. 
Savings was to focus on establishing and facilitating a voluntary savings program, modeled 
largely after CARE Zambia’s urban experiences and other successful savings programs. 
The savings component never evolved as a major activity in Mahavita. In the early phases of the 
project, under cash for work activities, there was a mandatory savings of 10% for cash-for-work 
participants. After their work tenure was completed they then received their savings in a lump￾sum payment. The activity was kept simple and participants were encouraged to use savings to 
invest in productive activities. For many the savings provided an important safety net, for others 
it was a cash ‘bonus’ they could use for consumption needs. When Mahavita was forced to start 
Table 3: Frequency and Percent of Household Members Trained in Various Fields
FFW Non FFW Comparison 
Profession Frequency 
Valid 
% Frequency 
Valid 
% Frequency 
Valid 
% Total
 
% 
Business 5 2 23 14 21 14 49 9
Industry 36 16 42 26 39 26 117 21
Agricultural Production 2 0.8 4 2 3 2 9 2
Artisanal Skills 18 8 17 10 16 11 51 9
Sewing, cooking, hair styling 24 10 37 23 36 24 97 18
Languages 4 2 12 7 8 5 24 4
Computers 2 0.8 10 6 14 9 26 5
Chauffeur, driver, etc 9 4 16 10 14 9 39 7
Maintenance of canals, dikes, 
etc. 131 57 2 1 1 1 134 25
Total 231 100 163 100 152 100 546 100
CARE Madagascar: Title II DAP Final Evaluation Page 13 
using food instead of cash in their public works activities the savings component was dropped. 
There had been intentions to revamp it under FFW, but this was never completed. 
The savings component was, however, integrated into the personal empowerment training. 
Despite the lack of a focused strategy for promoting savings there is some evidence that the 
benefits of having savings (briefly covered in personal empowerment trainings) were understood 
by households. In the baseline survey 36 percent of households reported having some type of 
savings. This increased significantly in the final survey to about 50 percent of households. Only 
about 35 percent of those households that were not residing in a project Fokotany and did not 
receive Mahavita training had savings, which was not any different than the percentages reported 
in from the baseline survey. FFW participants, however, had a lower savings rate than non-FFW 
participants. It is also significant to note that over half of those with savings reported having to 
withdraw funds within the previous year, and this was due primarily to the political crisis that hit 
the country. 
FFW Activities
A total of 16,892 households benefited from food for work activities during Mahavita, which 
was 169 percent of that planned. The reasons that the targets were exceeded were due, in part, to 
the political crisis that hit Madagascar in 2001, and CARE’s ability to leverage funding from 
other sources in responding to the crisis. 
Mahavita effectively and successfully selected the poorest and most vulnerable urban residents – 
one per household – as laborers to work on the food-for-work activities. Targeting success is 
supported by a number of indicators in the quantitative survey, such as assets, savings, and 
household expenditures (Figure 1). FFW laborers in those Fokotany’s visited by the evaluation 
team reported that FFW opportunities resulted in a significant short-term impact on household 
income, which translated into a positive short-term impact on household food security, providing 
a safety net for food insecure households for the period that they were in FFW. 
Residents interviewed during the evaluation agreed that the process for selection of FFW 
laborers was fair and transparent. Most of the workers heard about the job opportunity through 
their community Structures de Developpement or through ‘word of mouth.’ 
FFW had a positive impact on household diets. At the end of Mahavita, only 29 percent of 
general respondents thought that their diet was improving, but in FFW households, 40 percent 
felt that their diet was improving, with most noting an increase in meat and fish consumption or 
daily rice consumption and attributing this to their FFW participation. 
FFW laborers overwhelmingly expressed dissatisfaction with the content of the food basket, 
which consisted of maize, corn soy blend and oil. Nearly all FFW participants sold much of 
their maize to one local miller and used the proceeds to purchase rice. The transaction costs, 
however, were quite high. Normally they would sell the maize at about 1,000 FMG per kilogram 
and then immediately buy rice at between 1,800 and 2,000 FMG per kilogram. Although 
participants did not always understand how their pay rates were calculated, they appeared 
CARE Madagascar: Title II DAP Final Evaluation Page 14 
satisfied that their pay was fair, based on work output, and others were not profiting from their 
work. 
FFW workers normally put up with no more than minor scraps and scrapes from stones that 
occasionally hit people’s legs, arm or even eyes. CARE took out insurance on FFW laborers in 
case of illness during their participation. 
Figure 1: Expenditure patterns of FFW and non-FFW households. 
CARE took considerable precautions to responsibly implement FFW activities. In general, FFW 
was not used until community development structures were organized well enough to understand 
its purpose, take some management role, and use it towards sustainable activities. By and large 
it was used as a short-term solution for long-term development e.g. for initial canal cleaning, 
while regular canal maintenance would be covered by a community fund. 
IV.A.2 Strategic Objective 2 
Communities will be empowered to address constraints affecting their ability to 
control their own resources and livelihoods. 
Developing Community-based Structures
It has only been in the last decade that Madagascar has progressed to the point where democratic 
processes are allowed to freely operate at the community level. While this history is brief, there 
is a long history of patrimonial rule characterized by tight controls over access to resources and 
services. CARE took time to introduce processes that were culturally foreign, or at least 
politically sensitive not that long ago. Mahavita was a project primarily about developing 
representative structures at the community level, with the aim of raising the voices of the 
disenfranchised and catalyzing democratic processes within poor urban communities so that 
development could happen from within. In this respect, Mahavita made significant progress in 
an operating environment largely untested in these processes. 
Expenditures (%) - FFW HHs
73%
5%
6%
4%
3%
5% 4%
Food
Educat ion
Housing
Clot hin 
Healt h
Savings
Ot her
Expenditures (%) - General Household Sam ple
61%
9%
7%
4%
4%
9%
6%
Food
Educat ion
Housing
Clot hin 
Healt h
Savings
Ot her
CARE Madagascar: Title II DAP Final Evaluation Page 15 
Fokotany Selection
CARE Madagascar did an excellent job of targeting needy Fokotanys in Antananarivo. Of the 
30 Fokotanys covered by Mahavita, 55 percent were purely urban areas and 45 percent were 
mixed urban and rice production areas. TOUCH 2000 had worked in 10 Fokotanys and 
Mahavita continued to work in these areas and then added an additional twenty. The areas were 
chosen by Mahavita staff according to criteria that were developed collaboratively. The program 
began by first choosing three areas for pilot activities. Three was believed to be large enough to 
capture the diversity of the areas, while small enough to conduct a satisfactory study of the 
performance of program processes and procedures. 
Needs-based evidence suggests that the Fokotany’s selected by the project were indeed among 
the worst off. In terms of income, for example, quantitative surveys revealed that the thirty 
Fokotanys were among the poorest. Anthropometric data also showed that Mahavita fokotanys 
were worse off. For example, 49 percent of children under five are stunted in the project zone as 
opposed to 37 percent in comparison zones (other relatively poor Fokotanys that were not 
selected). Table 5 under Strategic Objective 4 provides further anthropometric evidence. 
Mahavita did not impose the project on any Fokotanys, opting instead to “invite” participation at 
the will of the existing local Fokotany administration. For the pilot phase Mahavita chose III G 
Hangar, Ankasina and Angarangarana because of their relative representativeness of the poor 
Fokotanys in general. III G Hangar is a completely urban area, and also the best performing 
Fokotany from the TOUCH 2000 project. Ankasina is a mixed urban and agricultural area and a 
medium performer of TOUCH 2000. Angarangarana is also a mixed urban and agricultural area, 
but it is new to the program. 
Empowering Structures de Developpement
One of Mahavita’s objectives was to create effective community-based structures empowered to 
identify, plan, implement, coordinate and sustain activities that would contribute to the well￾being of residents. These Structures de Developpement (SD) would form the core of the 
community empowerment process and would complement the personal empowerment processes 
of the project. The SDs were developed using wholly transparent processes and were based on 
what became known as FAMOA. 
“Mahavita worked within the rhythm of the community.” No better complement can be paid to a 
project that put poor people and the communities they live in at the forefront of its development 
efforts. It is impossible to overlook the tremendous progress and transformations that have taken 
place within communities as a result of explicit community empowerment processes fostered and 
nurtured by CARE and community members. The benefits of the empowerment processes have 
been realized by both individuals and communities at large. For perhaps the first time in their 
lives, community members have taken explicit charge of their own development, and have been 
willing to explore the parameters of community-driven development in partnership with CARE. 
It cannot be overstated what this means and the potential it holds for future development in these 
poor urban communities. 
CARE Madagascar: Title II DAP Final Evaluation Page 16 
In its first year, Mahavita worked with three Fokotany’s in a pilot phase that used participatory 
livelihood assessments as a basis for developing Representative Structures, the original label 
used for representative community-based organizations. Considerable freedom was allotted to 
Mahavita staff to explore how to apply participatory tools in engaging communities. Eventually 
the methods evolved into a process that Mahavita called FAMOA. This was a participatory 
planning process based on a partnership between communities and Mahavita and was used to 
facilitate the creation and strengthening of SDs7
 throughout the thirty Fokotanys. 
FAMOA began with an environmental analysis led by the community. It was used to validate 
and complement information obtained during initial livelihoods analyses. A Social Analysis 
followed and with information collected through focus groups. Information gained during the 
environmental and social analyses of the Participatory Livelihood Assessment provided a deep 
understanding of the livelihoods of different categories of households, their relative levels of 
livelihood security, and the principal constraints and opportunities to improving their livelihood 
security status. 
To date, SDs have been developed in all thirty of the target Fokotanys. The SDs were to include 
representatives of every organization and association in the Fokotany, as well as represent all of 
the social classes identified in the social analysis. Members of the SD are voluntary, and often 
are headed by an influential community member such as the chef de quartier. Because Mahavita 
‘followed the rhythm of the communities,’ there is considerable variation in the resulting SDs. 
There is no fixed membership, so in some SDs such as Ankor Andranomaheny there are seventy 
active members and thirteen bureau members and in Andavamamba Anjezika II there are 
twenty-one members. 
SD Activities and Management
A typical SD spent about one year in the FAMOA process and another 12-18 months 
formulating a Plan de Quartier (PDQ). This plan represents the analysis by the community of 
their priorities for development, and is the centerpiece of the FAMOA process. The 
development of the PDQ is also the glue that has held the SDs together and provided a common 
goal for everybody involved. Many PDQs reflect the communities desires in environmental 
infrastructure improvements, such as canal cleaning and pathway constructions, but can also 
include activities in beautification, community mobilization around health issues, and recreation 
facilities. 
It is interesting to note that SDs have, in most cases, evolved to being an unofficial arm of the 
official Fokotany structure. Fokotany officials are appointed by government, but because of the 
overt activities of the SDs, and perhaps also because of their transparency of operation, the SDs 
are embraced by the Fokotany officials and used sometimes by officials for a variety of purposes 
such as accessing food-for-work activities or soliciting assistance from outside agencies. 
There is absolutely no doubt that the creation of SDs and the empowerment of community 
members is viewed as an extraordinary and positive benefit to community members. 
 
7
 The term Structures de Developpement, or SD, replaced the term Representative Structure after the pilot phase. 
CARE Madagascar: Title II DAP Final Evaluation Page 17 
CARE has been very tactical in its work with Fokotany’s, and has established a rapport with 
community members unlike any other organization in Madagascar. Responses such as “CARE 
has been like a good brother” or “We have worked with other organizations and they do not treat 
us the same way as CARE” were common throughout the evaluation. There is a genuine respect 
between Mahavita staff and communities. 
What people point most often to, when asked about the benefits they have received from being 
involved with their SD, is a change in their own perspective about their community and their 
personal development. People tend to feel that they are part of an important process and that 
their efforts will eventually pay off in improved living conditions and opportunities. This 
outlook is supported by qualitative data gathered during the review and by quantitative data 
collected during the final survey. Table 4 shows how attitudes have changed with respect to how 
one’s opinions are perceived. In the baseline, just over 60 percent of community members felt 
their opinion was valued when it came to important community decisions. This increased to 83 
percent for FFW participants and 75 percent for non-FFW participants. 
Table 4: Perceptions on the value of community member opinions. 
Your opinion is valued when it comes to important decisions to be made in the community.
Household Type 
Baseline Final FFW Final non-FFW Comparison 
Agree entirely 60.5 83.5 74.4 70.5 
Do not entirely 
agree 
 8.3 6.3 12.6 8.8 
Disagree 26.4 6.8 9.7 15.9 
No opinion 4.8 3.4 3.4 4.9 
Figure 2 shows residents’ attitudes toward local infrastructure and perceptions on the quality of 
the local environment. Over 60 percent of FFW participants felt that infrastructure was 
improving, compared to about half of non-FFW participants and only 28 percent in non￾Mahavita Fokotanys. Similar trends were found related to perceptions of the local environment. 
Over 80 percent of FFW households felt the environment was cleaner. Almost 60 percent of 
non-FFW members felt the environment was cleaner compared to slightly less than 40 percent if 
comparison households. These data, along with other evidence in the quantitative report, 
strongly suggest that the work of SDs and infrastructure development are having a positive 
impact in the community, and that participation in activities such as FFW promote more positive 
outlooks by those participating. 
CARE Madagascar: Title II DAP Final Evaluation Page 18 
Perceptions on Neighborhood 
Environment
0
20
40
60
80
100
Cleaner No Changes Dirtier No Opinion
% of HHs
FFW HHs non-FFW HHs Comparison
Opinions on the Status of Local Infrastructure
0
10
20
30
40
50
60
70
Improving Staying the
same
Declining
% of HHs
FFW HHs
non-FFW HHs
Comparison HHs
 Figure 2: Resident Perceptions of Infrastructure and Environment 
 
IV.A.3 Strategic Objective 3 
Communities will create and maintain a healthy and hygienic environment within 
which inhabitants can carry out productive livelihoods. 
Infrastructure Development
Infrastructure provision started relatively late in the program, being preceded by several years of 
community organizing and development of the PDQs, which are a prerequisite to starting a 
community construction activity. Infrastructure is separated into “income generating” and 
“HIMO” or FFW. By far the greatest number of constructions have been in the HIMO, not 
income generating, category. Mahavita never intended to solve the infrastructure needs of the 30 
FKT, but to make reasonable headway in addressing these needs. 
Canal drainage and construction of pathways and “diguettes” (built-up walkways) have been the 
main focus of HIMO infrastructure improvement. About 60 kilometers of these pathways have 
been built, and there is no doubt that this reflects the priority of residents in the majority of the 
Fokotany’s where Mahavita works – most of which are located in the low-lying floodplain of 
Antananarivo. Every year the rainy season flooding causes extreme hardship for residents - 
relocation, resorting to travel via pirogue, exposed to highly contaminated water. Also, this type 
of construction is an ideal FFW activity. There is clear evidence of improved, cleaner 
neighborhoods thanks to this operation. Canals that used to be stagnant swamps now flow, with 
clean pathways linking different settlements. One woman expressed her delight at finally not 
having to move to the Catholic church every year during the rainy season. Much of the activity, 
however, has taken place since the last rainy season. The true “livability” of the improved 
neighborhoods will be tested in the next few months when the rains really start. 
There were other potential benefits to agriculture accrued by improving canal infrastructure and 
retention basins. Improved water management should have had a positive impact on agricultural 
production by improving water flow through canals, preventing prolonged or unplanned 
CARE Madagascar: Title II DAP Final Evaluation Page 19 
inundation, and improving drainage. However, there were no systematic efforts to assess any 
changes to agricultural production systems. 
Overall the quality of construction in Mahavita has been high. There have been some isolated 
problems with quality, such as inferior soil compaction on some dikes and walkways that may 
cause problems after heavy rains. There are also isolated cases of stagnant water collecting 
around wash basins and latrines due to poor leveling. These were pointed out to engineering 
staff and were to be corrected. 
Water/Sanitation/Washbasins
The “income-generating” category of infrastructure consists mostly of “blocs sanitaires,” a 
public toilet/shower facility usually with 2 of each, a caretaker who collects fees established by 
the management committee, improved public water taps, and public “basins lavoirs” or laundry 
facilities. These are all constructed in cooperation with AGETIPA, a government public works 
agency financed by the World Bank. A total of 37 have been built in 18 (of 30) neighborhoods, 
meeting construction targets. Of these 26 are functional, and 11 are waiting for water and 
electricity to be turned on. AGETIPA constructions are generally of high quality, the blocs 
sanitaires almost startling compared to the poorly built houses of the residents. The only 
exception is some problems with uneven land leveling around the structures, resulting in some 
collection of stagnant water. 
The investment per toilet seat seems very high compared to the need of the community for 
facilities. Various people estimated that over half the residents in the FKT visited did not have 
access to latrines or toilets, and small children rarely use any. Both Mahavita technicians and the 
SD selected the sites. One bloc was within 5 visible household latrines, so apparently only 
passers-by use it. 
Blocs sanitaires are the only sanitation or excreta disposal facility improvements carried out by 
Mahavita, and they were limited in number. Due to the participatory nature of Mahavita, 
whereby communities developed their own infrastructure priorities, no clear targets were set for 
sanitation improvements. No household or family-level excreta disposal facilities were built 
during DAP I despite the fact that only 15-17 percent of households have latrines in the poor 
neighborhoods, the majority of which are shared with other houses and 90% are outside. Some 
people have built their own latrines and charge a small fee for others to use them. A family 
latrine model adapted to the locale was developed and proposed to AGETIPA, but in the end 
they only built the official Commune model. A sanitation campaign exhorting each family to 
build a household latrine did not work. 
During walkabouts in the neighborhoods, latrines were pointed out (although one neighborhood 
claimed not to have any). In general, the superstructures were of very poor quality, dilapidated, 
some with large cracks in the slats so the person inside is visible, and often locked. The critical 
problem mentioned by ADCs and residents alike is the high water table of the zone, making it a 
technical challenge to dig pit latrines. The AGETIPA blocs sanitaires were constructed above 
ground to avoid flooding. A review of all PDQs developed by the FKT showed that nearly all 
CARE Madagascar: Title II DAP Final Evaluation Page 20 
proposed to build public “blocs sanitaires” but not family latrines or WCs, indicating either that 
these are not considered feasible or are not within the SD’s mandate. 
Water supply is still an issue in these zones, in spite of several new or improved water taps. 
Over 93 percent of households collect their water from outside their household. This 
necessitates that public taps be accessible and operational. In one of the newer FKT where 
Mahavita activities are barely a year old, the evaluation team was told that one unrehabilitated 
tap served 5,000 people, and only 20 percent of households have access to piped water. Even 
improved water taps have restricted hours of operation, although at the newer taps no lines were 
observed and people came at will to collect water during operating hours. 
All water collection observed was done by buckets, and both the baseline and final surveys 
showed that about 95 percent of water collection is by bucket (seau). Household water storage 
was not possible to see, but the follow up to the baseline survey claims that the majority of 
households (about 72%) use the same receptacle for storing both drinking and cleaning water, 
while about 27 percent use different containers for storage. In general, water quality in Tana has 
improved since the cholera scare of 2002. 
Sur Eau, the household chlorination product promoted by CARE and others, appears to be 
widely available but not widely affordable in the Mahavita FKT. Almost 98 percent of the 
respondents in the final survey had heard about sur eau as a product for treating water, which 
was up from 82 percent reported in the baseline survey. The primary source of information on 
sur eau was radio broadcasts, but many households also learned about the product and 
procedures from local organizers (animateur fokontany) and television. Conversations during 
community visits revealed that the use of sur eau is very low and associated with prevention of 
cholera, which is not a current threat. The final survey found only 15.2 percent of households 
making use of sur eau as a water purification method. This is a slight increase from the 13.7 
percent of households using the product during the baseline survey. People commented that sur 
eau was too costly for everyday use and that the water from the taps is clean enough for 
drinking. However, people also said that they were able to stem the cholera epidemic thanks to 
Sur Eau and are convinced of its efficacy against the disease. 
Despite the relatively low usage of sur eau in Mahavita households, it should be noted that it was 
the initial sur eau activities conducted under the DAP (initiated by CARE in partnership with 
PSI and CDC) in Antananarivo during the cholera crisis that led to the development of the sur 
eau model which is now national. Should future cholera or other water-borne disease epidemics 
emerge in Antananarivo the efforts of CARE under this DAP will likely be noted by surges in 
the use of sur eau for water treatment. 
Hygiene Promotion
Much has already been said about the remarkable community empowering process carried out by 
Mahavita. Combined with a thoughtful IEC strategy, results in terms of overall community 
cleanliness are visible. IEC on health, nutrition, water/sanitation, hygiene, and reproductive 
health is woven into every activity, and most of all into the HIMO/FFW activity where the 
majority of workers are women, direct channels to the poorest households. 
CARE Madagascar: Title II DAP Final Evaluation Page 21 
Mahavita’s impact in community-level hygiene improvement is perhaps most evident in the 
difference between “old” and “new” FKT. Older FKT - where there has been a CARE presence 
for up to 7 years - appear overall cleaner, with less garbage strewn about, than newer ones where 
CARE has been working for a year or so. This may be due to a variety of factors (such as the 
current “clean FKT” IEC campaign and competition), but casual conversation with residents 
revealed a consciousness of responsibility for the hygiene of the community due to CARE 
interventions through the SD. One effective hygiene promotion strategy is the training of 
community-based “animateurs” and “animatrices”, outreach workers who go door-to-door for 
hygiene promotion and counseling. They take their job seriously and are taken seriously by the 
residents. 
Knowledge of good hygiene behavior is very high in the program FKT. Themes, messages and 
graphics have been well-researched and tested. However, several people mentioned a disconnect 
between knowledge and practice. Several reasons were given for this. One, the Mahavita IEC 
strategy for behavior change has tended to focus on themes (i.e. “Nutrition”) rather than on 
specific behaviors. This was relatively recently identified as a problem within the IEC unit, and 
the Clean FKT Campaign was developed. It lasted for several months, and culminated in a 
festival community festival with prizes. A longer-term strategy for addressing non-adoption of 
hygiene behaviors does not yet exist. A second reason mentioned is poverty. People in the 
Mahavita zones are so desperately poor that they do not consider behavior change important. 
Another obstacle to adopting improved hygiene behaviors cited by interviewees has been the 
frequent problem of the lack of availability of the necessary hardware to carry out the promoted 
behavior. For example, the pervasive lack of latrines is a clear obstacle to safe disposal of feces. 
Hygiene promotion continued even when it was clearly not feasible for certain communities to 
do what was recommended. 
The problem of community garbage should be mentioned here, as it is an important contributing 
factor to disease (plague) in Antananarivo. It is a problem far from solved, although advances 
have been made, especially since the Commune instituted garbage collection from its “bacs”. 
The FFW efforts have undoubtedly resulted in an increased awareness of garbage as a 
community problem and a community responsibility. However, a number of people mentioned 
the problem of the “bacs” being located far from the center of a FKT (near an accessible road, 
for instance) and the unwillingness of residents to carting their household waste to the “bac”. 
Much of it thus still ends in the street. 
Social marketing has centered on promoting the use of Sur Eau. As mentioned before, it was an 
effective program and mitigated the cholera outbreak, but is not consistently applied. Cost was 
mentioned as a factor. It did, however, permit Mahavita and others to gain a better 
understanding of beliefs and behavior around water use. 
Enabling Environment
The solid community foundation created by Mahavita through the FAMOA process and 
embodied by the Structures de Developpement and their various committees has been discussed 
CARE Madagascar: Title II DAP Final Evaluation Page 22 
and appreciated throughout the evaluation. It has also been pointed out that there is uncertainty 
among the 30 SD as to how to implement the community projects in their PDQ, and many of 
these projects are directly linked to the environmental health of the community. The SDs visited 
were among the most mature, and seemed nearly - but not quite - ready for handling the 
negotiations needed with the Commune and ministries with whom the PDQ would be executed. 
It’s doubtful that the SDs with less maturity will be able to manage without continuing capacity 
building from CARE. 
The enabling environment for environmental health at the community level should relate to the 
management of water/sanitation and related systems (basins lavoirs) to assure availability and 
quality of services. Mahavita’s process for building SD capacity to manage cost recovery fits 
into its overall community empowerment strategy and involves the creation and training of a 
management committee who conducts a willingness to pay study of their own community and 
establishes user fees based on the results. This system appeared to be functioning well, with 
carefully filled-in ledgers and in a few cases, filled coffers with the prospect of cash on hand for 
the SD. It also bears mentioning that nine community groups received municipal contracts to 
clean canals after having cleaned their own. 
Another enabling factor for improved EH is Mahavita itself and its capacity to assure the 
availability of required materials and personnel for carrying out a program of infrastructure 
improvement and capacity building. It appears as though the separate Mahavita sectors 
responsible for the technical aspects of infrastructure improvement and for community 
mobilization strategies had some gaps in coordination, resulting in occasional communities 
mobilized and ready for work, but supplies and materials undelivered. 
In spite of good faith efforts throughout the Mahavita program to coordinate and collaborate with 
the Commune of Antananarivo and technical agencies responsible for service provision in the 
FKT, this proved to be extremely difficult and frustrating. The result is that much of Mahavita’s 
activities were done independently. The future of the relationship between the SDs and the 
Commune, and the need to advocate during the next DAP for official recognition by the 
municipality of the SDs is the subject of the paper on future directions for Mahavita. This is 
especially critical for assuring the environmental and other health of the residents of the FKT 
since in a dense urban environment, especially in a flood prone zone crisscrossed by drainage 
canals, a coordinated effort for environmental improvements is required to assure coverage. 
Fecal contamination and rats don’t recognize a FKT boundary.
CARE Madagascar: Title II DAP Final Evaluation Page 23 
IV.A.4 Strategic Objective 4
Households practice healthy behavior and hygiene.
Health Sector
Mahavita’s health unit focuses its activities mainly on taking care of the health of the FFW 
workers, support to the IEC division in the development of technically accurate health messages, 
and is launching a pilot HIV/AIDS prevention activity. It has virtually no connection with the 
official health sector even though the program has its roots in child health under TOUCH 2000. 
Meetings with several MOH representatives at various levels were warm and interviewees 
expressed a strong desire to renew ties with CARE in its urban work. 
Reasons for the separation between CARE/Mahavita and the health sector are numerous and 
mainly involve policies and personalities from the previous GOM administration, as well as a 
conscious move by Mahavita to focus on environmental health as the critical priority of the 
target FKT and move away from direct health interventions. The MOH officials interviewed all 
felt that CARE’s strength in community mobilization and the current MOH priority strategies 
were a good fit, especially in poor urban areas where health service needs are great and success 
of some MOH interventions depends on community activism. Immunization and TB treatment 
were cited as examples.
Despite the lack of focused, direct health interventions there is information from the monitoring 
component of Mahavita related to health practices. Illness is common and about half of all 
households will have had at least one family member ill within the previous two weeks. Of those 
households with illnesses about one-third are seeking medical treatment. Table 5 provides a 
breakdown of the different illnesses experienced by households. The most common malady is 
acute respiratory illness (ARI), accounting for one-third of reported illnesses, followed by 
localized aches, and flu. 
Table 5: Incidences of selected illnesses from the Mahavita final survey.
Flu Diarrhea ARI Aches Vanishing Skin Other 
n % n % n % n % n % n % n % 
General HHs 84 21.7 41 10.6 128 33.1 99 25.5 4 1.0 6 1.6 25 6.5 
FFW HHs 29 23.0 22 17.6 40 31.7 24 19.0 1 0.8 1 0.8 9 7.1 
Non-FFW HHs 28 21.8 12 9.4 43 33.7 34 26.6 1 0.8 3 2.3 7 5.5 
Comparison HHs 27 20.3 7 5.3 45 33.8 41 32.0 2 1.5 2 1.5 8 6.8 
In general, knowledge of appropriate health behaviors is high, even among the urban poor. For 
example, over 85 percent of households with children under two years of age have had their 
children vaccinated. Knowledge of diarrheal causes is very high, as well as treatment. 
Malnutrition
As shown in Table 6, height for age measurements of children below five years of age suggest 
serious chronic nutrition problems continue in Antananarivo’s Fokotany’s, resulting in high rates 
of stunting. Children from Mahavita FFW households had the highest rates of stunting at over 
CARE Madagascar: Title II DAP Final Evaluation Page 24 
55 percent, while non-FFW households were not significantly different, with stunting rates at 48 
percent. Comparison households had the lowest stunting rates of 39 percent. These rates are, 
however, significantly better than those found during the beginning of Mahavita. During the 
baseline survey (Table 6), almost 64 percent of the urban population of children below five years 
of age was found to be stunted (Z = < -2.0 standard deviations). No causality was sought in the 
baseline survey, but its results combined with other studies in Madagascar suggested that diet 
diversity and general health status of children were two main contributors to stunting. The 
stunting rates found during the final survey are still high, especially for children residing in the 
project area. 
Table 6: Anthropometric data for children under 5. 
HH Type Measurement n 
% Below –3 std 
deviation 
% Below –2 
std deviations Mean 
Height for Age 151 22.5 55.6 -2.111 
FFW Weight for Height 151 0.6 7.3 -0.728 
Weight for Age 151 10.6 47.7 -1.842 
 
Height for Age 103 22.3 48.5 -2.067 
Non FFW Weight for Height 103 1.0 5.0 -0.511 
Weight for Age 103 3.2 36.9 -1.702 
 
Height for Age 89 12.4 39.3 -1.783 
Comparison Weight for Height 89 1.1 9.0 -0.842 
Weight for Age 89 10.1 40.5 -1.754 
 
Height for Age 186 29.6 63.6 -2.363 
Baseline Weight for Height 186 0.5 2.0 -0.072 
Weight for Age 186 3.8 35.7 -1.523 
 
CARE Madagascar: Title II DAP Final Evaluation Page 25 
IV.B CYPREP (Cyclone Preparation on the East Coast) 
Special Strategic Objective: To safeguard the livelihood security of Madagascar’s most 
cyclone-vulnerable communities by enhancing local capacity to plan for and cope with cyclone 
emergencies. 
The evaluation of the CYPREP component was conducted against its three intermediate results 
and including the March 2001 readjustment. For each IR multiple strategic orientations were 
developed - four were designed to contribute to the achievement of IR1, two to IR2, and one to 
IR3. The various evaluation themes are organized and presented below with the appropriate 
strategic orientations. 
IV.B.1 Intermediate Result 1: Reduced vulnerability in 132 communities on the East 
Coast of Madagascar. 
The project met its expectations for this result, working in 133 communities out of a planned 
132. The intervention sites were divided as shown in Table 7. 
Table 7: Geographic location of CYPREP intervention sites 
Zone Sub-prefecture / District Number of 
sites % Total by zone % 
Southern Mahanoro 25 18.8 25 18.8 
Northern Fénérive-Est 25 18.8 60 45.1 
Vavatenina 17 12.8 
Soanierana Ivongo 11 8.3 
Tamatave II 7 5.3 
Central Tamatave II 24 18.0 48 36.1 
Brickaville 24 18.0 
 TOTAL 133 100.0 133 100.0 
In terms of location by zone, the northern zone had the most sites (45.1%), the southern zone had 
36.1%, and the southern zone had 18.8% of the intervention sites. 
Strategic Orientation 1: Involve local authorities throughout the process of site selection. 
From the outset CYPREP made strong attempts to involve local authorities in the community 
selection process. Several steps were used to selection sites for the CYPREP component. First, 
a workshop was held bringing together the sub-prefect, communal authorities and mayors, and 
project staff. The objectives and strategies of CYPREP were discussed and future plans were 
developed. Next, mini-workshops were held in each commune selected, following a letter of 
interest received from communal authorities. In addition to CYPREP orientation, these 
workshops incorporated an overview of the livelihood and disaster preparedness situation in each 
village. Village representatives (PCLS – fokontany, representatives of the fokonolona, and 
village elders) presented the situation in their respective fokontany, and at the end of the meeting 
the assembly established a prioritized list of potential intervention sites to guide future processes. 
 
CARE Madagascar: Title II DAP Final Evaluation Page 26 
Finally, the project field agents (AMS) conducted a preliminary site study of likely villages, 
using the prioritized list prepared by each assembly as a guide. This included collecting baseline 
information to supplement the presentations prepared by the village representatives, with a 
special emphasis on evaluating the situation of the most vulnerable elements of the population. 
Using participatory processes resulted in the following criteria used to select intervention sites: 
- Dynamism of the community and the local leaders (PCLS, village elders); 
- Accessibility of the site - in general sites that were not more than 2 hours walk by foot from 
the nearest motorcycle access point were given higher priority; 
- Population: fokontanys of more than 1,000 residents were given priority; 
- Willingness of the community to participate in the planning, management and execution of 
cyclone preparedness and mitigation activities. 
The selection process allowed for the active participation of local authorities in the choice of 
sites, despite some evidence of subjective bias by certain elected officials. From the point of 
view of the local authorities the site selection process, as well as the criteria used, was 
appropriate and well adapted to the local context. However, the use of accessibility as criteria 
handicapped certain vulnerable communities located in more remote areas, despite their high 
level of motivation and interest. According to certain local leaders, the site selection process 
also allowed each commune to choose villages that were representative of the situation in the 
commune as a whole. According to some of the project staff, the project should have also 
considered the specific conditions prevailing in each zone, notably the access/transport situation, 
to make the eventual implementation of the project more feasible. 
The manifestation of interest in the form of a written letter on the part of the community was a 
very pertinent indicator in the selection process. The system used to present the situation in each 
fokontany resulted in the development of knowledge, motivation, and capacity for the village 
representatives. 
The important role played by the local authorities in promoting and facilitating the project can be 
linked to the fact that site selection was not a unilateral decision made by project staff, but rather 
a joint decision between project staff, the local authorities, and the community representatives; 
CYPREP also had a requirement that each community provide a 20 percent contribution on all 
mitigation activities with the exception of road rehabilitation. It required communities to collect 
resources from all households over time. This stipulation may have intimidated certain 
communities in the beginning, especially in areas where there had recently been cyclones (as was 
the case for portions of Mahanoro and Brickaville), and it may have excluded the most 
vulnerable communities that were too poor to meet the criteria. 
In addition to site selection, local authorities, administrative as well as traditional, were involved 
in the prioritization of mitigation activities and in determining the mix of community funds and 
in-kind contributions to be collected. The community members interviewed in the course of the 
evaluation stated that they were satisfied with this process. Local authorities also participated in 
CARE Madagascar: Title II DAP Final Evaluation Page 27 
the Conseil Villageois de Secours (CVS), the deciding body for planning, analysis, and decision￾making for the implementation of development activities in the village. 
Strategic Orientation 2: Use the household livelihood security analysis approach (HLS) to 
design development activities to prevent and mitigate the adverse effects of cyclones on target 
communities; reinforce the capacity of local communities to identify, plan and design 
development activities. 
The FALAFA Process (HLS)
Between 1999 and 2000, the project concentrated its efforts on resource inventories and risk 
evaluations, with a goal of developing cyclone preparedness plans (PPC). This effort was 
somewhat ad hoc and lacked a strong conceptual framework to guide information collection and 
analysis. After the strategic reorientation in 2001, Mahavita staff helped CYPREP adopt the 
FALAFA process to better facilitate the development of village development plans (PDF) in the 
intervention sites. This was a conceptual framework based on a Household Livelihood Security 
(HLS) approach and adapted to the local context. 
Because the two approaches (risk analysis leading to a cyclone preparedness plan and HLS 
analysis leading to a development plan) were introduced at different points in the evolution of 
the project, the field agents tended to view them as separate processes, even in sites where both 
the PDF and the PPC were completed after 2001. The process permitted the following results to 
occur: 
- Mitigate the vulnerability of the community; 
- Give the community a better understanding of cyclones and their effects; 
- Identify and choose the mitigation activities best corresponding to the situation and capacity 
of each community; 
- Invent / conceive new mitigation strategies based on local traditions and resources; 
- Facilitate the creation of a local body / committee charged with emergency preparedness and 
relief; 
- Develop a more comprehensive view of the community; 
- Measure the degree of motivation and vulnerability in the community; 
- Increase the motivation and dynamism of communities; 
- Bring a new spirit of openness and participatory decision-making; 
- Simplify the identification of problems and risks; 
- Identify the root causes and links between problems, thus giving the community the 
possibility to make more informed choices about the solutions to implement; 
- Put in place a profound and evolving system of community analysis. 
For the local authorities, the process allowed them to clarify and better identify the needs of the 
community, especially as related to development. However, the process was sometimes too 
long for the community, and often due to time constraints different community members would 
participate in different sessions, sometimes creating a discontinuity in the analysis. 
CARE Madagascar: Title II DAP Final Evaluation Page 28 
The benefits of FALAFA as an information collection and analysis approach were significant. 
At the top of the list would be the level of community involvement in reflecting on their own 
livelihoods and its relation to disaster management. This in itself led to improved 
communication within the villages. The creation of a forum for discussion and reflection 
(CVS) created a nexus for catalyzing action in the community and improving the social stature 
and knowledge of CVS members. An increase in participation of women in the development 
process was also noted, as was improvement in analytic ability and technical skills, especially 
among younger CVS members. 
The processes used also resulted in better knowledge of the level of vulnerability of each 
community, and permitted community members, especially the CVS, to identify the different 
roadblocks to development in their community and to devise appropriate solutions. Many local 
authorities noted their satisfaction with the benefits derived from using the FALAFA process and 
they would like to see it extended to other villages so that they, too, can play a more active role 
in local development. 
Fokontany Development Plans (PDF)
CYPREP facilitated the development of village development plans (PDF) in each community. 
The PDF is a summary document that includes an analysis of problems, resources, and 
opportunities and the proposed solutions and actions to be undertaken for the development of the 
community. The plan is developed by the community with the assistance of one or more project 
field agents) using the FALAFA framework. Out of a total of 133 sites where the project worked 
89 communities (67 percent) had developed a PDF by the time of the evaluation. 
The development of a PDF in each site was one of the most interesting and beneficial results of 
the CYPREP project, both from the point of view of project staff as well as the communities 
themselves and the communes. The PDF represents an indispensable tool for all development 
actions to be undertaken in the community, irregardless of subject matter. 
In general, the CVSs and community members seem satisfied with their PDFs and the processes 
used to develop them. Many expressed appreciation of the opportunity to participate and the 
knowledge they gained. Most also feel that the activities outlined in the PDF are realistic and 
respond directly to the needs identified by their community. Results from the quantitative survey 
suggest that most community members are aware of the major themes of their PDF and are 
satisfied with its contents. All those interviewed were unanimous in stating that the PDF is an 
indispensable tool for the development of the community, which is indicative of a sense of 
ownership on the part of the local authorities, the members of the CVS, and the community at 
large. Some communities have already used their PDF to either create or revise PCDs 
(commune development plans). The existence of the PDFs helps the elected officials at the 
commune level to identify the needs and wishes of the various fokontany (villages) and to assist 
them in advocating for the fulfillment of them; 
Many of the early sites in Fénérive Est, which were targeted by CYPREP between 1999 and 
2000, do not have PDFs. However, the abbreviated risk analysis conducted there has permitted 
CARE Madagascar: Title II DAP Final Evaluation Page 29 
many of these communities to proceed to an analysis of problems and risks in the context of 
emergency preparedness. 
In the final survey, 54% of the households surveyed in the intervention sites were aware of the 
existence and general contents of the PDF (almost all the households in the sites where a PDF 
was completed). This compares with households surveyed in non-intervention sites, where most 
were unaware of the existence of PDFs or confused with the PCDs (commune level development 
plans). 
Cyclone Preparedness Plans (PPCs)
The cyclone preparedness plan (PPC) is conceived by the community with the assistance of a 
project field agent. Based on the local disaster context, its aim is the attenuation of the negative 
impacts of future cyclones. The PPC contains a list of the actions to be undertaken before, 
during and after the passage of a cyclone. Of the 133 CYPREP intervention sites, all but one has 
completed a PPC. 
Just over 60 percent of the households surveyed were aware of the PPC, compared with 5 
percent of the households surveyed in non-intervention sites (whose communities do not have 
PPCs). The creation of the PPCs in Malagasy appeared to foster a heightened sense of 
responsibility on the part of community members, reinforced by the creation of sector 
commissions and the training of the members of the CVS. Many of the households surveyed in 
intervention villages were also able to recite the contents of the PPC. 
As a result of CYPREP activities it appears that communities are better prepared to respond to 
cyclone events. Some examples of important behavior changes includes the use of cyclone flags, 
warnings broadcast by the dalala8
 and goka9
 and understood by community members (other 
types of morse-style conventions are also known and understood by the community); protection 
of food stocks and clean water in safe locations; reinforcement of buildings, removal of tree 
branches likely to fall on buildings, etc. In general, there is much more understanding by 
community members concerning the impact of cyclones and the action needed to mitigate them. 
In terms of knowledge of households about alert systems, signal systems, and weather alerts, the 
intervention villages consistently showed higher knowledge compared to non-intervention sites. 
The difference in level of knowledge ranged from a maximum of 91 percentage points on one 
indicator to 7 percentage points on the lowest. Three-quarters of the population in the 
intervention villages (77 percent of households surveyed) know the alert system used in their 
village, compared with only 10 percent in non-intervention sites). A similar proportion was 
aware of a coordinating body (i.e., the CVS) that is responsible for cyclone mitigation activities. 
Almost all the households surveyed in the intervention villages (90%) stated that they had 
learned about the alert systems from the members of the CVS. Knowledge of the cyclone flag 
was shown by about three-quarters of the population in intervention villages (77%), compared 
 
8
 Dalala: vocal calls 
9
 Goka: conch shell used as a horn 
CARE Madagascar: Title II DAP Final Evaluation Page 30 
with 28% in non-intervention sites. The difference in knowledge between the intervention and 
non-intervention sites averaged 74 percentage points in Mahanoro. In the other districts the 
distinction was less pronounced with an average difference of 44 to 45 percentage points for all 
indicators. This could be explained in several ways, including the recent passage of cyclones in 
Mahanoro, allowing the intervention communities ‘real-life’ practice of their new skills and 
knowledge. Also, project interventions in Mahanoro began later than in other zones, meaning 
that, on the average, less time had elapsed between the initial training of CVS members and the 
evaluation. Finally, the baseline level of the non-intervention communities was lower in 
Mahanoro, perhaps due to the paucity of other development actors in that district. 
The project produced a surprising level of spread effect in non-intervention villages. With 
respect to the cyclone flag, one alert system known by a significant percentage of respondents 
(28%) in non-intervention villages, 56 percent had heard of it from CVS members from a 
CYPREP site while 40 percent heard about it from the radio. 
The IEC activities undertaken by the project, through the intermediary of the CVS, had a positive 
impact, as almost every household in the intervention villages had undertaken cyclone 
preparation measures outlined in the PPC. From this, one can conclude that the vulnerability of 
these households has been reduced. As an example, according to the communities of Vohiboazo 
and Andovoranto, they were able to avoid loss of life from Cylone Manou (May 2003), whereas 
previous cyclones had always resulted in at least one death. According to the mayor of the 
commune of Tamatave Sub-Urbaine, some villages, which didn’t work with CYPREP, have also 
developed PPC (e.g., Vohibolo). This was a product of exchanges between neighboring villages 
and the CVS members trained by the project. 
The post-cyclone evaluation of damage and loss (filling out of EIMA forms and sending them to 
the CCS or CLS) was completed by all CVSs in the CYPREP intervention sites within 48 hours, 
whereas most non-CYPREP villages had not been able to complete the evaluation in under a 
week. Besides their timeliness, the information provided by the CVSs was deemed reliable and 
useful for planning the emergency interventions of the CNS. 
Strategic Orientation 3: Design cyclone mitigation activities based on HLS analysis and ensure 
the sustainability of these actions by reinforcing the management capacity of local communities. 
Mitigation Activities
The mitigation activities initiated with the assistance of CYPREP contribute to reducing the 
negative impacts of cyclones. Major mitigation activities are those that can not be entirely auto￾financed by the community themselves. The actions that can be undertaken by the community, 
without outside assistance, are auto-mitigation activities: reinforcement of buildings, 
construction of family latrines, trimming of overhanging tree limbs, etc… 
The major mitigation activities conducted over the course of the project included: rehabilitation 
of rural roads, cyclone-resistance community granaries, and cyclone shelters, rehabilitation of 
CARE Madagascar: Title II DAP Final Evaluation Page 31 
drainage and irrigation systems, potable water points, motorized pirogues, and reforestation of 
watersheds. The levels of effort of these different activities are provided in Table 8. 
Table 8: CYPREP results with respect to major mitigation activities. 
Rehabilitation Reforestation(nb) 
Road 
Irrigation 
and 
Drainage 
Canals 
surface 
Shelter Granary Cloves Litchi Cinnamon Eucalyp Total 
Boat 
Potable 
Water 
Zone Point 
(km) (km) (ha) nmbr (m²) nmbr capacity 
(t) nmbr nmbr nmbr nmbr nmbr nmbr nmbr 
North 182 0 0 25 1 388 23 93 19 018 5 810 2 746 58 867 86 441 2 14
South 65 26 610 2 0 4 0 7 158 200 0 0 7 358 01
Center 28 3 40 1 40 0 0 0 0 0 0 0 01
TOTAL 274 29 650 28 1 428 27 93 26 176 6 010 2 746 58 867 93 799 2 16
Out of a total of 274 kilometers of rural road rehabilitated, two-thirds were located in the 
northern zone, 24% in the southern zone, and 10% in the central zone. In terms of reforestation, 
out of a total of 93,799 saplings, 92 percent were planted in the Fénérive-Est zone and only 8 
percent in Mahanoro. Tamatave did not have any reforestation interventions. 
IV.B.2 Intermediate Result 2: Creating a network of local partners for disaster 
response. 
Strategic Orientation 4: Development of a monitoring system to track CYPREP and provide 
information to the Network of Rural Sentinel Sites and livelihoods analysis. 
Strategic Orientation 5 was designed to provide timely information to CYPREP staff on the 
progress of activities and to apply the results of livelihood inquiries into activity programming. 
The monitoring system, however, was not fully functional until five months before the end of the 
project, which obviously impeded its impact on activities and programming. 
The monitoring system was based on periodic reports from each of the three zones. However, 
there was not a formal monitoring and evaluation unit within CYPREP, and no one individual 
was charged with collecting, analyzing and reporting on information. Rather, it was the 
responsibility of all project staff to report on and apply information. The Senior Management 
Team in Antananarivo had a difficult time following project progress due to the lack of 
systematization in the monitoring system, and often decisions were made based on less-than￾optimal information. For all it was difficult to apply real-time lessons learned because of the 
status of the monitoring system. 
The monitoring system was also intended to feed information into the Network of Rural Sentinel 
Sites and to assess and affirm CYPREP’s understanding of livelihoods. Its tardiness highlighted 
above prevented this from being realized. 
CARE Madagascar: Title II DAP Final Evaluation Page 32 
Strategic Orientation 5: Development of more representative community organizations capable 
of coordinating and managing their development activities. 
A total of 131 Village Security Committees (CVS) worked with CYPREP, corresponding to a 
total population of 190,000 inhabitants. Only one site in Mahanoro and another site in Fénérive￾Est do not currently have a CVS or any mitigation activities as a result of CYPREP’s work. 
Table 9 summarizes CVSs by zone. 
 Table 9: Existence of CVSs by zone. 
Zone Number of CVS Number of CVS 
members Total population 
Fénérive-Est 60 1,534 88,431 
Mahanoro 25 854 38,686 
Tamatave 48 1,536 62,866 
Total 133 3,924 189,983 
The CVS is a village-level coordinating body charged with overseeing the planning and 
implementation of development activities and preparing for and mitigating against natural 
disasters. Officials from different administrative levels (sous-préfet, mayoral, chef de quartier) 
as well as community representatives unanimously agree on the usefulness of the CVS for 
disaster-related matters as well as for community development. 
The positive benefits that have been noted by participants and as a result of CARE’s establishing 
and supporting CVSs are many. First and foremost, it is the only community-based structure on 
the east coast that supports local development based on community representation and 
participation. 
CVSs have evolved as the community mobilization force around which shared values and 
visions are being discussed and acted upon. They have played an important role in making 
community members aware of their roles and responsibilities related to cyclone preparation and 
response, and in changing behavior of community members with respect to cyclone preparation. 
As a result, the arrival of a cyclone is no longer unforeseen by communities and the risks to the 
general population have been diminished. The CVS also benefits the commune and the CCS by 
providing a direct linkage to communities, households and individuals. 
The CVS is known and recognized by approximately 75 percent of the population residing in the 
CYPREP intervention zone. The activities most well known by community members that are the 
responsibility of the CVS include cyclone warnings and communications of impending events 
(recognized as a CVS function by 73 percent of residents polled), the construction of disaster￾related infrastructure (33 %) and livelihoods/asset protection (31 %). 
After the passage of a cyclone, the CVS is responsible for promoting health messages to 
communities to prevent illnesses to assure safety of the population. This responsibility is 
recognized by 34 percent of residents. Fewer recognize the CVS as being responsible for 
CARE Madagascar: Title II DAP Final Evaluation Page 33 
reconstruction (20 %) and other health-related activities (19 %). The vast majority of these 
activities are not known in villages where CYPREP has not been active and there is no village￾level structure for disaster management. 
Accomplishment of CVS Tasks 
The high motivation of CVS members is one of the key factors in making it an important 
development and disaster-related community body. The In the majority of cases there are no 
major problems in its functioning. The sole exception noted during the evaluation was the rare 
event where certain individuals could not accept consensus-based decisions and the general 
functioning of representative structures. 
Despite the overall success that has been achieved with regard to CVSs there remains work to be 
done. The CVS could benefit from becoming a recognized legal structure whose responsibilities 
are to provide preparedness and mitigation measures related to disaster management and tom 
promote community-based development. Indeed, all evidence points to endorsing the CVS as 
legal structure so that it is justly recognized by authorities at all levels. More active engagement 
in disaster planning by local authorities would ensure better linkages to communities and well as 
create a more dynamic and integrated disaster preparedness and response capability throughout 
the east coast of Madagascar. It would also reinforce the motivation of CVS members. One 
other factor that has helped to keep CVS motivation high is the use of small ‘rewards’ to CVS 
members such as CVS hats, training certificates, t-shirts, etc. 
IV.B.3 Intermediate Result 3: Re-enforce the capacity of the National Emergency 
Preparedness and Management Agency (CNS) to promote the use of 
preparedness planning and disaster response measures at the local level. 
The Conseil National de Secours (CNS), the government’s agency responsible for planning and 
coordinating disaster relief interventions, is part of the Ministry of the Interior. Its historical role 
in disaster management has been limited due to a lack of resources. In most cases, its structures 
exist only in the form of lists of government officials who are responsible for responding to an 
emergency on behalf of their technical Ministry. Both UNICEF and UNDP have worked 
directly with CNS at the national and regional levels. UNICEF’s strategy in particular has aimed 
at training regional representatives about their roles and responsibilities in case of emergencies. 
These efforts are crucial, however there are only six regions in Madagascar, and it has long been 
evident that regional CNS representatives, even if they are motivated and clear on their purpose, 
cannot single-handedly manage disaster preparedness and response for the whole of their 
populations without some assistance. 
CYPREP’s aim was to fill the gap between the CNS as a coordinating body and communities, 
which are ultimately the victims of disasters. Until CYPREP, no planned action for disaster 
management had been undertaken at the village or community level. CYPREP’s assistance in 
the development of a network of organized village level disaster committees created the means 
by which the CNS could be both influenced by and support the communities for which they are 
responsible. CARE has tried to play a key role in enhancing the institutional capacity of CNS in 
CARE Madagascar: Title II DAP Final Evaluation Page 34 
building up its information systems and strengthening its role as national coordinator of disaster 
response activities. 
The results have been mixed, primarily due to the difficulties of working with regional bodies 
and due to the community-based workload undertaken by CYPREP. It is apparent that 
preparedness and response strategies elaborated by each community will eventually help CNS 
play its role, but CYPREP has been challenged in doing both local-level development and 
regional capacity-building at the same time. Timing is better now for focusing on strengthening 
the CNS now that the communal and community linkages have been established by CCSs and 
CVSs. 
Other parallel efforts by CARE to strengthen CNS, however, have been made. Since 1999 
CARE Madagascar has orchestrated a system called SIRCAT for the development of an 
Information System for Risk and Disaster Management within the CNS. This system has 
succeeded in gaining national consensus on the geography of hazards and vulnerability in 
Madagascar. 
Strategic Orientation 6: Support for the development and operations of the Commune 
Emergency Management Committee.
The Commune Emergency Management Committee
 
The Commune Emergency Management Committee, or CCS, was established in all CYPREP 
communes and some communes outside of the projects focus. All received training in disaster 
preparedness and mitigation by CNS with CYPREP assistance, but the majority of CCSs are not 
very functional. Their members are ad hoc, and many still do not know their roles and 
responsibilities in relation to either the functioning of the CCS or disaster response. It should be 
noted, however, that CYPREP is only now in a phase of providing assistance to CCSs in, for 
example, elaborating Communal Plans for Cyclone Preparation (PPCC), and that this has 
occurred at such a late stage in the DAP that it has not had a chance to be properly evaluated. 
Capacity-building of Social Mobilization Agents (Assistants de Mobilization Sociale (AMS))
All Social Mobilization Agents (AMS) recruited and used in the project received formal training 
at the beginning of their service on social mobilization, household livelihood security, and the 
FALAFA process. This type and level of training was an important experience for AMS agents, 
all of whom come from rural areas in the CYPREP zone. Despite their limited exposure to the 
training topics they were able to grasp key concepts and apply them in their work. 
In each zone it was CYPREP’s policy to put into place a multi-disciplinary team in order to 
facilitate team work and provide a broad perspective to community development and disaster 
management. 
Despite CYPREP’s efforts to provide high-quality training, several AMSs cited areas where they 
felt their capacity needed more strengthening. Part of the reason for this may be the broad range 
of issues that they face when working in rural and often remote areas. Many of these 
CARE Madagascar: Title II DAP Final Evaluation Page 35 
communities have had only minimal exposure and interaction with development organizations, 
and often the AMSs are overwhelmed. Some of the suggestions made were to enhance the 
training to include communication skills, rights-based approaches, children’s and women’s 
rights, pedagogy and health issues. Many also expressed in receiving computer training. Other 
suggestions included having exchanges among AMSs both intra- and inter-zonal in order to 
reinforce training, exchange experiences, and develop common approaches. All of these 
capacity-building elements would, they felt, enhance their ability to facilitate the FALAFA 
process and assist communities in the development of their PDF. 
IV.C Food Aid Management Systems & Internal Controls 
A full accounting and evaluation of CARE’s food aid management and practices was not called 
for in the Scope of Work for the final evaluation of Mahavita. However, some observations are 
provided here that resulted from the evaluation work. 
CARE Madagascar has one large warehouse in Tamatave with a capacity of 1,500 tons and two 
prefab warehouses (e.g., 350 tons in Fenerive) to store the food commodities. Together they have 
a capacity to store about 2,000 MT. The food was stacked properly and on pallets in clean 
conditions. CARE Madagascar has occasionally experienced commodity infestation, but was 
able to work closely with the government to salvage infested commodities. Monthly physical 
inventory counts were periodically reconciled with stock balances. In all, it appears that proper 
warehousing management has been practiced in CARE Madagascar. 
CARE Madagascar has followed the Food Resource Manual Guidelines for commodity 
management. Designated personnel receive and issue commodities upon authorization of the 
Project Coordinator. Waybills were properly prepared for receipts and issues of commodities. 
Inventory reports were prepared monthly and submitted to CARE’s head office for checking and 
reporting to the donors. 
In addition, the finance section of the project controlled food transactions. Food Monitors 
engaged in random and independent checks of the beneficiary lists and activities accomplished. 
The quality of the food distribution system was monitored and included end-user monitoring. 
IV.D Food Distributions in FFW Programs
CARE Madagascar distributed food for emergency and regular programs based on emergency 
and FFW ration scales developed within the project. CARE prepared food distribution plans for 
the whole year and submitted reports by comparing the plan to the actual food deliveries. Some 
beneficiaries could not tell the team about the quantity and type of food rations they received. 
However, they believe the food distributions were fair and equitable. FFW participants noted a 
short-term improvement in household food security. Because the communities controlled the 
beneficiary selection process, targeting for emergency programs was usually deemed to be fair. 
CARE Madagascar: Title II DAP Final Evaluation Page 36 
V. Conclusions 
V.A Mahavita 
Strategic Objective 1: Targeted households will be using increased income and savings to 
provide for their food and livelihood needs. 
ƒ The FFW activities in Mahavita were run with the utmost professionalism, and as a result 
relatively few problems were encountered. The poorest Fokotanys and poorest households 
within Fokotanys were all well-targeted as evidenced by qualitative and quantitative data. 
Female participation was also high. Target benchmarks in FFW activities were exceeded by 
about 70 percent. 
ƒ Mahavita was able to leverage donor resources during the political crisis and rapidly initiate 
FFW activities that provided a vital safety net to poor urban households. This rapid response 
capability was the direct result of working with developing Structures de Developpement and 
the rapport that Mahavita had built with Fokotany residents and officials. The actions taken 
during this time period validated CARE’s approach to community development and the 
impact that creating representative structures can have on mobilizing financial and human 
resources. 
ƒ Poor communications among Mahavita, USAID Madagascar and FFP Washington on 
resolving the CFW-FFW issue created delays in programming and restricted some 
alternatives that may have been better than those implemented. The amount of time it took to 
resolve the issue also hindered progress and may stalled momentum that had been building 
through the CFW activities. 
ƒ Too little was done to generate employment opportunities for Fokotany residents during the 
life of the project. Use of cash-for-work and then food-for-work for employment generation 
was not a sustainable employment generation activity. Rather, it was an important safety net 
for poor urban households, especially during the political crisis. FFW is not a substitute for 
employment generation. Mahavita did not have a viable strategy for generating employment 
opportunities but depended on an early hypothesis that significant demand could be 
generated for the services of worker groups that gained new skills under FFW programs. It 
appears that no other alternative strategies were ever employed to transfer skills gained from 
FFW and personal empowerment training to viable income-generating activities. 
ƒ An opportunity was lost on carrying the savings component over when CFW was shifted to 
FFW. The savings component was an explicit objective of the project but was virtually 
ignored after the cash-for-work activities were transferred to food-for-work, even though 
there are viable models for establishing savings schemes in FFW activities. 
CARE Madagascar: Title II DAP Final Evaluation Page 37 
Strategic Objective 2: Communities will be empowered to address constraints affecting their 
ability to control their own resources and livelihoods. 
ƒ CARE clearly had no well-articulated exit strategy for disengaging with SDs. The 
development of the PDQ became the de facto end product of the community development 
process, and Mahavita staff struggled internally with how to “keep the PDQ alive.” It would 
have been beneficial to pilot strategies to implement the PDQs once the earlier and more 
mature SDs finished their plans. Some SDs have had their plans finished for over one year 
without guidance on how to seek funding or how to promote their local plans in other ways. 
ƒ All of the SDs are struggling with ‘next steps.’ Members still appear highly motivated, but 
this motivation could quickly wane if the PDQ does not result in tangible outputs and if there 
are no further activities to keep members engaged. The PDQ should be viewed as being 
closer to the starting point of the community-development process than to the end point of 
the process. Some SDs are seeking their own funding with limited success, but there has 
been no coordinated effort on the part of Mahavita to develop alternative strategies and 
capacity for such action. 
ƒ Mahavita’s strategy of obtaining Fokotany approval of the project through an official request 
and naming the core Fokotany authorities as the initial ad hoc committee was effective. This 
is a valuable lesson learned for working in urban areas where initially there may be little 
community cohesion and where it is desirable to avoid the frequent mistake of using 
infrastructure as an entry point. 
ƒ In many ways Mahavita became stalled on the success of its FAMOA process. The 
achievements in developing participatory structures at the community level were so great that 
staff ignored some of the other vital components of livelihood enhancement strategies, 
namely income generation and savings. 
ƒ The political crisis had both positive and negative impacts on the project. On the positive 
side, as noted above, CARE was able to leverage donor resources during the political crisis 
and rapidly initiate FFW activities that provided a vital safety net to poor urban households. 
This was a result of the SDs already being in place and capable of rapidly organizing their 
communities. CARE was able to respond much quicker than any other NGOs. On the 
negative side, the crisis diverted attention away from the natural evolution of project 
activities, and is likely the main reason why more advancement was not made on 
implementing PDQs. 
ƒ Often dominated by the chef de quartier or other influential person. Difficult to ascertain 
how ‘representative’ the SD is over time, as there is little idea of constituency and almost no 
rotation of those who ‘manage’ the SD. 
ƒ More cross-fertilization of SD development may have been a good strategy to generate 
lessons learned and provide SD members with a sense of camaraderie among their neighbors. 
Also a forum where SD members meet to discuss advances, problems, successes, etc. This 
CARE Madagascar: Title II DAP Final Evaluation Page 38 
forum would have been a good entry point for brainstorming new ideas or building consensus 
among Fokotany members. 
ƒ The legal status of the SDs and their positioning in light of recent GOM administrative 
changes will need to be addressed in order to determine what type of Fokotany structure 
CARE will continue to support in future urban work. For example, there may be scope for 
small government subsidies to employers who hire local residents. 
Strategic Objective 3: Communities will create and maintain a healthy and hygienic 
environment within which inhabitants can carry out productive livelihoods. 
ƒ CARE/Mahavita has made some serious positive inroads in the environmental health of an 
urban zone that is extremely poor and vulnerable. FKT where program activities have been 
ongoing for a number of years clearly have healthier, more hygienic environments. 
Residents speak of noticeable health improvements (less skin disease, children’s diarrhea, 
even infant mortality). These results do not show, however, in newer FKT, and CARE is 
challenged to continue its work in the original 30 FKT and to add 50 more, all with reduced 
personnel. 
ƒ CARE/Mahavita has workable models for community organizing and empowerment, social 
mobilization, cost recovery, and for linking infrastructure development to food security and 
poverty reduction. However, most of the infrastructure is too new to yield models for 
maintenance. 
ƒ CARE’s strategy of avoiding building infrastructure before the requisite community-driven 
development process is completed is sound. Past experience in other urban projects has 
shown that when infrastructure is focused on first then ownership issues arise and 
maintenance can be a real problem. However, Mahavita probably started too late in the 
project on infrastructure, with a significant amount being developed during the last six 
months of the project. This makes it difficult to ascertain the extent to which communities 
will maintain the infrastructure. Most SDs have maintenance plans and/or committees, but 
most of the infrastructure is too new to determine how well it will be maintained. 
ƒ In spite of the fact that CARE/Mahavita attained the majority of its construction targets (the 
exception being with latrines), water and sanitation coverage of the target FKT is inadequate 
for the needs of these populations. This is especially the case for sanitation (safe feces 
disposal), and lacking this, it is difficult to say that communities have completely achieved 
healthy and hygienic environments. Despite the success of the health and hygiene 
components in promoting knowledge and behavior change, it is difficult for some households 
to take full advantage of their knowledge due to access issues associated with latrines. The 
AGETIPA model does not seem appropriate for meeting the sanitation needs of this 
population and future interventions in sanitation infrastructure need to reevaluate latrine 
design and strategy. 
CARE Madagascar: Title II DAP Final Evaluation Page 39 
ƒ Certain themes or community environmental health issues have not been adequately 
addressed. In particular, food safety in the common gargottes and hotely where the baseline 
survey found over 50% of the population eats every day, and hand washing with soap. These 
themes are part of the package of themes and messages, but haven’t been spotlighted. 
ƒ Decisions concerning environmental health strategies such as HIMO and promotion of Sur 
Eau were logically made within the political and epidemiological context of certain periods 
(food emergencies, political upheaval, and cholera). Things have changed and calmed down, 
and the kind of strategic rethinking required by the new DAP is timely. 
ƒ CARE’s use of GIS technology was innovative and had a positive impact on its credibility 
with the communal government. Internal issues on managing the GIS system stalled its 
usage in latter stages of the project. 
Strategic Objective 4: Households practice healthy behavior and hygiene. 
ƒ A systematic household level evaluation was not possible. However, the follow-up to the 
Baseline has substantial information regarding a range of household hygiene behaviors: 
water availability transport and storage, latrine availability and use. It also contains 
information on health status and practices: incidence of diarrhea and ARI, care seeking in 
case of child illness, children’s nutritional status. 
ƒ Very unscientific observations and discussions gave the following impressions: many 
households in older FKT have a new consciousness concerning their responsibility for 
maintaining their home environment (sweeping, garbage disposal). However, the lack of 
usable latrines makes it doubtful whether there can really be healthy, hygienic households. 
While the follow-up to the baseline shows a high percentage of latrines in the zone, it was 
repeatedly mentioned as a problem during the evaluation. A shift in focus to adequate 
sanitation coverage is critical. 
V.B East Coast Programming - CYPREP 
CYPREP had only one strategic objective, thus the conclusions are organized around its three 
Intermediate Results. 
Intermediate Result 1: Reduced vulnerability in 132 communities on the East Coast of 
Madagascar. 
ƒ CYPREP would have benefited from prioritizing the rehabilitation and reopening of rural 
roads earlier on in the intervention process (assuming that in each case it was also a 
community priority), in order to put the potential intervention sites on more even footing. 
This would have led to better targeting of the most vulnerable communities. However, it is 
recognized that road rehabilitation is an ambitious undertaking and the limits imposed by 
CARE Madagascar: Title II DAP Final Evaluation Page 40 
obligatory environmental studies would have made it difficult to entirely eliminate the 
criteria of accessibility from the site selection process. 
ƒ While CYPREP staff did conduct site visits and surveys afterwards to validate the 
information presented in the workshops and provide a more realistic image of the situation, it 
would have been advantageous to collect this information before the workshops, so that 
discrepancies between the various information sources could be discussed in the workshop, 
and allowing the village representatives to benefit from the information collected. One 
possibility would be to develop a village-level data base for the region. This would allow a 
more objective discussion of indicators and vulnerability from the start of the process, and 
would eliminate time spent verifying biased data. What is more, if the local authorities were 
also implicated in the creation of the data base, they themselves would have better access to 
objective information about their communities. On the other hand, the time and resources 
needed to conduct an exhaustive survey of all fokontany would have been prohibitive and 
possibly delayed the initial project implementation. A third option would be to conduct the 
fokontany-level survey only in the communes pre-selected in the district workshop – a 
compromise which would not require such extensive time and resources. 
ƒ A framework for the presentation of each fokontany was prepared by CYPREP and used by 
village representatives during the mini-workshops. This should be expanded, to improve the 
quality of the information presented by the local authorities and to assist those who are 
handicapped by their low level of education. Some flexibility should be included to account 
for the varying levels of education of local authorities. This would eventual lead to the 
development of a standard tool for presenting information at the village level. 
ƒ CYPREP did not develop any type of consensus-based standards for defining vulnerability. 
Without some form of standards, and information systems that complement these standards, 
it is impossible to determine if vulnerability has changed and how. The absence of a strong 
monitoring component to CYPREP was perhaps its largest shortcoming, and it failed to take 
full advantage of post-disaster opportunities to critically assess its impact and use reflective 
practice to inform its programming and strategy. 
ƒ CYPREP’s requirement that each community contribute 20 percent of the necessary 
resources/funds for major mitigation activities was too strict. Future community contribution 
schemes should be more flexible and include different forms of contribution, both monetary 
and non-monetary. For example, there could be phased contribution to allow work to start 
with a community commitment to invest. This would be a “mature” community that had 
participated in other project activities and had demonstrated previous community 
contributions. A combination of monetary and physical contribution could be used for 
poorer communities and could be needs-based or applied generally across the community. 
Or a scaled contribution according to a project-based definition of vulnerability could be 
used. 
CARE Madagascar: Title II DAP Final Evaluation Page 41 
Intermediate Result 2: Creating a network of local partners for disaster response. 
ƒ The participation of community members in prioritizing mitigation activities as well as 
deciding the level of community contributions was one of the strong points of the approach 
used by CYPREP. This aspect of the approach should be continued and expanded in future 
projects to increase community ownership. The same is true of the participation of local 
authorities, which should be continued and reinforced to help them better assume their 
responsibilities and those of the CVS. 
ƒ Despite the effort made by the project from the beginning to solicit the participation of the 
commune-level authorities in the different steps of the process, certain mayors suggested that 
it would have been even more fruitful to fully integrate the staff from the mayor’s office in 
the facilitation and training processes, so that they could better support and monitor the 
activities eventually undertaken by the villages. 
ƒ The writing of the PDFs in French is useful for the local authorities when dealing with 
donors and regional authorities. However, it is not useful to the community members at large, 
few of whom speak French. In recognition of the fact that the community is the end user of 
the PDF, the document should be developed entirely in Malagasy, with the French version 
being simply a synthesis of the most important points. This is essential if one wants to speak 
of full ownership of the document by the community. In this way, more members of the 
community could read the original document and discuss its contents (rather than relying on 
the verbal description of the contents given them by community leaders). This would also 
reinforce the use of the PDF as a tool for negotiation and reference amongst the different 
community groups. 
ƒ Because the PDF is an indispensable tool for the creation of a PCD, the active participation 
of local officials, especially the mayor’s office, in the planning process for each village 
should be continually reinforced. This will help local officials better understand their 
development role and will build communication links between them and their communities. 
ƒ As in Mahavita, CYPREP lacked an exit strategy for disengaging with communities, and the 
PDF was often viewed as the endpoint. At a minimum, a system for periodically updating 
the PDF and adding new information should be developed in future programming. This 
would also serve as an opportunity to retrain CVS members and reinforce their skills through 
experiential learning. 
ƒ Given the effectiveness of PPCs in changing practices related to disaster preparedness and 
mitigation it is important to find a way to develop PPCs in all East Coast communities. Also, 
the use of the Malagasy language contributes extensively to ownership of these plans by the 
community. Expanding coverage could be realized through the intermediary of the CCS, 
which should theoretically pick up the baton and extended the CYPREP activities throughout 
the commune. The training given to the members of the CCS, in all three intervention zones, 
conducted by CYPREP in collaboration with CNS and the United Nations Volunteers, will 
certainly contribute to building the capacity of the communes in this regard. 
CARE Madagascar: Title II DAP Final Evaluation Page 42 
ƒ It is also important to ensure that all villages have alert systems (gokas, flags, etc) and that 
there is a common code for the different types of cyclone alerts. This could be done at 
almost no cost, for all the non-intervention villages in the zone. It would perhaps be 
preferable, compared with providing the village with megaphones, which will eventually 
need repair or maintenance. Similarly, it will be necessary for CNS to provide the 
communes with adequate resources so that they can accomplish the tasks assigned to them, 
and to reinforce the capacity of the members of the CCS. 
ƒ The choice of mitigation options may not have fully been explored, and was often limited to 
(or influenced by?) the technical expertise of CARE staff, especially engineers. Future 
mitigation activities should explore the full range of livelihood options that relate to 
preparing for and mitigating the impact of disasters. 
Intermediate Result 3: Re-enforce the capacity of the National Emergency Preparedness 
and Management Agency (CNS) to promote the use of preparedness planning and disaster 
response measures at the local level 
ƒ Community members, local authorities and project staff were all unanimous regarding the 
usefulness of the CVS in reducing vulnerability of east coast households and in its efforts to 
mitigate the negative impacts of cyclones in particular and disasters in general. The CVS has 
evolved into an indispensable structure for all community-level development and in terms of 
cyclone prevention. In this regard, the development of such CVSs or equivalent structures 
should be regarded as a key element of any current or future strategy for east coast 
programming. une telle structure serait indispensable dans toute intervention au niveau du 
fokontany, car elle constituera la structure de relève lors du retrait du projet. De part son rôle, 
elle constitue : 
ƒ Eventually the issue of CVS legality must be addressed more directly. If the CVS is the 
community-based organization that guides local development and manages disaster 
preparedness and mitigation then it has to have a more permanent and respected status. This 
would require that the CVS is recognized as a legal structure by the GOM, that it has certain 
roles and responsibilities that are commonly and legally defined as well as recognized by the 
community, and that its active members retain a sense of commitment and purpose. 
It should be noted here that the same issue exists for SDs in urban programming, and CARE 
needs to develop a unified strategy for fokotany structure that it can advocate for in the 
future. 
ƒ The use of multidisciplinary teams in its programming is a very strong, positive element of 
CYPREP, and its system merits being copied in other programming. The dynamism within 
the CYPREP team was high and there was a feeling of empowerment among staff that was 
facilitated, no doubt, by the team processes that were established and supported (similar to 
staff feelings in Mahavita). 
ƒ The FALAFA process contributed to empowerment feelings among staff and AMSs. It 
would also be useful to supplement the FALAFA experience with training in 
CARE Madagascar: Title II DAP Final Evaluation Page 43 
communications and PRA skills, and with training techniques for adult literacy and 
andragogy. 
V.B.1 Other CYPREP Conclusions 
ƒ The monitoring and evaluation system was put into place too late in the project for it to have 
a meaningful impact on progress tracking and real-time feedback. Overall there was a lack 
of structured monitoring and learning. For example, the impact of Cyclone Kesiny (and 
others) on project activities was never fully researched, and in a disaster-related project often 
the only opportunity to learn about approaches and their impact on targeted communities is to 
do intensive post-crisis follow-up. 
ƒ CYPREP was hampered by not conducting a proper baseline. The original monitoring plan 
called for assessing temporal change of selected indicators based on their baseline values. 
Since there was no baseline this was not possible and made it difficult to determine the 
magnitude of some changes that took place. Of perhaps more importance would have been 
the experienced and knowledge gained by project personnel participating in the baseline 
survey. 
ƒ Future east coast activities should consider using a Management Matrix for monitoring 
institutional change. The Management Matrix is a tool composed of individual indicators 
organized around the themes of institutional strengthening and expected performance of 
institutions. Themes can include such items as governance, financial management, human 
resources, vision and planning capacity, and implementation. Under each topic is a 
negotiated set of indicators that the project and each institution (for example, CVSs) agrees 
upon. The indicators reflect shared but specific criteria for success in capacity-building. 
Indicators can be qualitative or quantitative and each has negotiated criteria scoring (the 
same indicator used for different institutions, thus, can have different scoring criteria). 
Scores can be aggregated within themes and for the entire matrix. There is a maximum 
obtainable score that can be used to assess overall capacity. 
ƒ Intensive implementation (multiple activities/interventions in one location) versus spreading 
out. Is there a “critical mass” for reducing cyclone vulnerability? 
VI. Recommendations for DAP II 
This current evaluation, as well as important work by Drinkwater (2003) and Garrett (2003), has 
identified important issues that need to be considered as CARE continues its urban programming 
efforts under the new DAP and other funding. The issues and recommendations highlighted will 
be incorporated into this section for ease of reference since their views are widely shared by the 
evaluation team. 
CARE Madagascar: Title II DAP Final Evaluation Page 44 
Urban Development/Governance
y Seek legal formation of SDs in order that they be recognized as the official development 
entity of the Fokotany. In doing so, engage Arrondissement and Communal officials in the 
process so that they recognize from the beginning the logic and value (as many already do) of 
such action. 
y Forge a rationale relationship between the newly elected representatives of the Fokotanys and 
the SDs. Explore with community members what the nature of these relationships should be, 
along with those of appointed officials. Facilitate the articulation of a more formalized 
arrangement for SDs. 
y Develop a clear exit strategy for the next DAP. The exit strategy should include when and 
how CARE will disengage from their work with specific Fokotanys and what the nature of a 
broader relationship would be. 
y Develop a set of alternative strategies for linking communities and their PDQs with service 
providers and sources of funding. There may not be a “one size fits all” approach to this, and 
the issue of the number of Fokotanys seeking these linkages should be considered, so that 
one alternative strategy may include the grouping of Fokotanys according to certain criteria 
(such as linkages to specific canals). Also, the formal status of the PDQs needs to be 
clarified, and advocacy is needed to make other institutions aware of its evolution, its linkage 
to community-based processes and structures, and its role in community-driven development. 
y Continue to use GIS as a tool for planning infrastructure development and maintenance and 
as a capacity-building tool for communities and government officials. Mapping of canals in 
the current DAP was an effective way to gain access to communal government and provided 
an important service for Antananarivo. 
 
y The issues of representativeness and constituency need to be revisited with respect to SD 
 function. Most SDs have had little or no turnover in leadership since their creation. This 
was a slight advantage to Mahavita during this DAP because it added continuity to a new and 
sometimes difficult process. In the longer run, however, more democratic processes should 
be encouraged and could even include periodic elections for SD leaders. The idea of SD 
leaders or members having constituents should also be explored. Since the SD has emerged 
as the ‘voice’ of the community there needs to be a reality check to gauge just how 
representative they are, and who speaks for different social groups that live in the Fokotany. 
y In the new DAP Mahavita staff will have to develop indirect models to provide support to 
SD. Using local NGOs is one option. Another is to use mature SDs to help develop 
capacities in new Fokontanys. It might be feasible for SDs to work with local NGOs, so that 
the longer term support is provided by the NGOs, and the time demands on the SD members 
are not too huge. Nevertheless, such a role is when for which the SDs should receive some 
compensation, perhaps for stages of the process evaluated to be successfully completed. 
Thus, in the development of a strategy to expand to new Fokontany, existing SDs, as well as 
CARE Madagascar: Title II DAP Final Evaluation Page 45 
relevant local NGOs, provide the major opportunities. The provision of joint training, to 
ensure SD/NGO partnerships, would be ideal. 
y In transitioning to the next DAP, Mahavita will need to make strategic decisions on levels 
and modes of operation with all urban stakeholders, in particular with service government 
authorities, service providers and donors. 
y The reduced Mahavita staff will have to forge strengthened relationships with a number of 
institutional actors for DAP II. Mahavita will need to develop an extensive relationship with 
the CUA and the Arrondissements and a good working relationship with service providers 
such as Jirama and Apipa. With still others, such as UNDP and some INGOs, CARE may be 
a collaborator in promoting urban governance. Additionally, sectoral decisions will need to 
be made – how much does Mahavita engage with the health sector institutions, for example, 
from whom a strong willingness to work again with CARE has been expressed to the 
environmental health consultant on the evaluation team. Since Mahavita cannot simply 
impose its own agendas, it will have to seek agreements with the institutions themselves, and 
how Mahavita can support these. In any case, some type of MOU with each organization
which specifies joint purposes and respective roles in promoting improved governance and 
service delivery within the fokontany would be a valuable starting point. 
y The new DAP will need to facilitate the planning process at the CUA/ arrondissement level, 
and this will require a coordinating body with active participation. CARE may chose to 
rehabilitate the ISIKA model. If so, it needs to address the sustainability problems that 
plagued previous efforts. In the Kuyakana project in Maputo a coordinating and planning 
body was facilitated by the expedient mechanism of the project paying the salary of a person 
acting as coordination secretariat in the municipal council. This role subsequently became 
institutionalized and the person graduated to becoming a full time council employee, once the 
value of the role had been clearly demonstrated. Initially the role focused on the 
establishment of regular meetings to coordinate the activities of different agencies 
responding to the aftermath of floods in Maputo. What is required is a purpose that would be 
widely accepted by different agencies that would serve as a means of reaching agreement as 
to how coordination could take place, and how the responsibility would be managed. 
• Federating SDs, for example, at the arrondissement level. There are far too many SDs – some 
192 in the CUA as a whole – for each to be able to negotiate resources for the 
implementation of their PDQs. There needs to be a much more strategic mechanism whereby 
the SDs can engage with stakeholders, and this can only happen with mechanisms for 
grouping the SDs. This could happen in different ways. Perhaps the most obvious would be 
for the SDs in each arrondissement to develop a federation, so they can work directly with 
the arrondissement and with other authorities, at that level. Such a structure could also liaise 
with bodies like Apipa. To begin with though, both the levels of federation of SDs and the 
mechanisms they could use to reach agreement across fokontany on priorities, would need to 
be agreed. And this in turn, would require some more formal acceptance of the SDs. 
• Another generic activity successfully utilized by Kuyakana was to hold periodic stakeholder 
fora on specific urban themes, many of them related to environmental health. The project 
CARE Madagascar: Title II DAP Final Evaluation Page 46 
would meet with a range of stakeholders beforehand, generate an understanding of their own 
agendas and challenges, and then use the one day stakeholder workshops as a way of 
bringing the different stakeholders together, have them share operational issues, and then 
agree on follow up steps and mechanisms. Some form of stakeholder engagement and 
collaborative mechanisms will be required for specific thematic issues. 
• In shifting to the next phase, and an approach that is based more on rights than needs, one 
critical role for Mahavita is to be able to work at the municipal and, where necessary, 
national level, to agree service provision entitlements for fokontany residents, particularly in 
relation to basic service provision. Standards, if they do not exist already, should be agreed 
for areas such as water provision, sanitation, drainage, and refuse removal, and then more 
critically, plans for ensuring that these levels of provision are met, should be negotiated. This 
role may well require advocacy work, and would be best undertaken if Mahavita is part of a 
broader coalition that agrees to pursue such changes. If necessary, Mahavita should be able to 
play a leadership role in the establishment of such a coalition. One role of the coalition would 
be to push for greater resource allocation in areas that are necessary, and where appropriate, 
the tracking of resources that have been allocated, to ensure they are used for their intended 
purposes. 
• One issue for Mahavita in this next phase will be the new staffing configuration and skills
that are needed, given that the program is about to change fundamentally the way it works. 
This applies particularly with respect to the various kinds of partnership work. Present staff 
will need to upgrade their skills and become more confident working at different levels of 
government and other organizations. Some new staff may also be required, if certain skills 
appear obviously lacking. GIS is one area, and advocacy skills may be another, though 
existing staff should be encouraged to take on this kind of role as a first step. In addition, 
another area where all staff require retooling, is their ability to shift to more of a rights based 
approach. This is likely to require external support. 
• Mahavita – and urban government in Madagascar – remains relatively closed to urban 
experience elsewhere, and it would help the program and staff to receive wider exposure. The 
use of judicious cross-visits could be extremely helpful to the program, particularly where 
urban programs are already working in a more indirect mode. Urban programs that have solid 
experience of supply side work and facilitating the ‘negotiated development’ role, would be 
most appropriate. The Shahar program in Bangladesh, or some of the urban programs in 
Zambia, could be good candidates. Some of the effective experience in working at the level 
of municipal government has been that of the Kuyakana project in Mozambique, but this 
project is currently seeking renewed funding and has only a skeletal operation remaining. If 
cross visits were to occur, they would be most appropriate if they occurred once better 
operational relations had been developed with the CUA and arrondissements, so that at least 
one staff person from communal government could also take part. In addition, as wells as 
potential cross-visits, other kinds of networking and communication should also be instituted 
that results in the broader exposure of Mahavita and CUA staff to urban governance 
experience in other relevant contexts. 
Environmental Health
CARE Madagascar: Title II DAP Final Evaluation Page 47 
y Focus on achieving better water/sanitation coverage in the target FKT by developing single 
or multi family latrine models adapted to flood prone areas. Collaborate with the WASH 
Campaign that is embarking on a program of operations research into such latrine models, 
but does not currently have a specific urban slum focus. 
y Develop a hygiene promotion strategy to generate demand for sanitation improvements, and 
support the Commune technical agencies responsible for providing these services. An 
overall promotion strategy could be piloted in several Fokotanys and then adapted as a more 
general approach. 
y Restart hygiene promotion (IEC) programs with a focus on key household behaviors such as 
safe excreta disposal, protection of drinking water, and hand-washing with soap. Coordinate 
closely with the hardware side such as family latrine construction. Even though current 
evidence suggest that health and hygiene knowledge is high, there is needed improvement in 
actual practices. 
y Expand the techniques for hygiene promotion to avoid stagnation and boredom. Link with 
others in and out of Tana/Mad who are doing similar activities, for exchanges and 
inspiration. Consider traditional theater, videos etc. 
y Strengthen the system of community “animateurs” and “animatrices” by increasing their 
capacity as community health agents through more training that groups different FKT, 
exposure visits to areas where results are being felt. Consider remuneration (as some already 
are). 
y Streamline the FAMOA and PDQ development process without sacrificing its effectiveness. 
This has been mentioned elsewhere. 
y CARE needs to be concerned about all aspects of the work it does in coordination with other 
bodies (e.g., quality of infrastructure, appropriateness of design) because ultimately it also 
reflects on CARE. For DAP II CARE should establish operating principles or principles of 
engagement with other donors or agencies to address this issue. Care helped AGITEP with 
World-Bank financed public latrines but their appropriateness and prospects for cost￾recovery are questionable. 
Health
y Consider reconnecting with the Health Sector at both the Ministry and the Commune levels 
(not the same) in mutually beneficial activities. In particular, consider collaborating in areas 
requiring strong community participation such as identifying dropouts from immunization 
program and from TB treatment program, organizing immunization campaigns, identifying 
malnourished children for treatment at specialized centers. Mahavita’s community 
animateurs/trices could play a critical role in these and other areas. 
CARE Madagascar: Title II DAP Final Evaluation Page 48 
y Work with the MOH to advocate for an urban slum policy. This policy has existed for some 
years in draft form, but the MOH and UNICEF haven’t been able to push it through. The 
time may be right for such a policy. 
y Support the official MOH business plan and strategic objectives and avoid parallel health 
activities. This is one reason there was estrangement between CARE and the MOH in 
previous years. 
Guiding Principles for DAP II Environmental Health activities
y Whenever and wherever possible, connect environmental health activities to income 
generation. For example, design a family latrine coverage strategy so that local artisans can 
make and sell parts, or that a caretaker paid by the families maintains multi-family latrines. 
Or, help communities organize contracted house-to-house garbage collection in areas where 
the “bac” is far from the residences. 
y Conduct all environmental health improvement activities with the cognizant Commune 
authorities, aiming at increased GOM awareness of the needs of the urban poor, and 
ultimately to improved urban poor policies and programs. 
y Seek out partnerships to support various priority programs not covered by the DAP. For 
instance: 
 - Link with JSI/RTI’s new program to support Communes to develop PDD leading to 
grants at the CSB level; 
 - Explore collaboration with WASH campaign on a number of fronts – latrine 
development, hygiene message and methods development, linking with others 
carrying out similar programs, especially during the upcoming Hygiene Promotion 
“Braderie”; 
 - Coordinate with other NGOs etc. with programs in areas of relevance to CARE 
population (housing, agriculture, HIV); and 
 - Develop an effective stakeholder and service provider coordination mechanism such 
as an Urban Health Alliance or an Arrondissement-level coordination model. CARE 
can play the role of convener of such a group. 
Enhancing Food and Livelihood Security on the East Coast
y In CYPREP, the involvement of local authorities in site selection contributed to the success 
and sustainability of the activities undertaken. For this reason, the involvement of 
authorities, administrative (sub-prefecture, commune, fokontany/village) as well as 
traditional (tangalamena), throughout the process is indispensable and should be continued in 
future projects. 
CARE Madagascar: Title II DAP Final Evaluation Page 49 
ƒ As a priority for future interventions, it would be profitable to assist the communities that 
began working with CYPREP in 1999 and 2000, in Fénérive Est, to also develop PDFs. This 
would assist future projects working with these communities, and would require very little 
effort, as the communities, in the course of their collaboration with CYPREP, already 
completed most of the baseline work. 
ƒ The communal officials would like to have a more comprehensive packet of training and IEC 
materials to help them in training villages that did not work with CYPREP. The communes 
that are in the process of creating their PPCC (communal cyclone preparedness plan) have 
completed PPCs for all the villages. 
ƒ It is undeniable that the participation of women in community-based activities brings a new 
vision to local development. As a result, it is important to consolidate, continue and expand 
the lessons learned concerning the integration of women in the development process, an issue 
on which the project has accumulated significant experience, permitting them to contribute 
more fully to the development of their communities. 
ƒ The emergency response of CNS, after a cyclone, is based on the report compiled by the 
CLS, which is in turn based on the EIMA forms arriving from each village. For this reason, 
the existence of a functional CVS, with members trained to fill out the form and convey it to 
the proper authorities, is a considerable advantage for CNS and for other actors offering 
emergency assistance after a cyclone. This CYPREP activity, in particular, merits being 
expanded throughout the region, to increase the rapidity and effectiveness of future cyclone 
response efforts. 
CARE Madagascar: Title II DAP Final Evaluation Page 50 
 ANNEX A 
Scaling up Community-Driven Development: 
Reflections on the Mahavita Experience10
James Garrett 
International Food Policy Research Institute 
IFPRI Reference No. 2570-023 
This report highlights key issues surrounding Mahavita’s operations and plans to scale up based 
on my visit to Madagascar from January 28 to February 12, 2003. These reflections are based 
on wide-ranging conversations with current and former senior and field-level staff, with 
municipal, prefecture, and fokontany authorities, and with members of the community structure 
de developpement (Annex 1). Andri Ratsitohara, head of follow up and evaluation for 
Mahavita, and I spoke with representatives of 10 different communities (out of 30 where CARE 
now works), 15 field staff, and 5 senior staff using semi-structured interviews, along with 
municipal and national authorities. (See Annex 2 for study methods, objectives, and interview 
guides.) 
This study will form the basis of one of a number of case studies that IFPRI is carrying out in 
collaboration with the World Bank on scaling up community-driven development. That case 
study will compare and contrast the Mahavita experience with that of PUSH / PROSPECT, 
another program initiated by CARE focusing on urban livelihoods, in Zambia. 
Although this report focuses on factors affecting scaling-up, it will also touch on operational 
issues. As the prefet de Antananarivo pointed out, “the two are linked. If (questions about 
operations) are not answered, it is hard to scale up.” 
Scaling up often refers to expansion in terms of area or numbers (that is, size), but it can also 
refer to adding new activities, including activities to increase political influence (Dongier et al. 
2002, Uvin 1995). Although this report will not focus on these distinctions, they are present in 
the Mahavita experience. 
The focus on scaling up is especially relevant given Mahavita’s ambitions to expand from 30 to 
80 Fokotany’s in the next DAP. The main expression of community-driven development (CDD) 
under Mahavita is the FAMOA process – the community-driven preparation of a community 
development plan, or plan de quartier, a PDQ. CDD does not mean that the community takes 
on full responsibility for all development activities in the area; but it does mean that the 
community participates meaningfully, and fundamentally drives, the decisions about 
development that affect it. 
In simplified schematic terms, for CDD to be effective, communities must have the capacity to 
carry out development-related activities and make demands on service providers, primarily 
 
10 A complete report, including annexes, can be requested from IFPRI or CARE Madagascar. 
CARE Madagascar: Title II DAP Final Evaluation Page 51 
government authorities. Government must have the willingness and capacity to respond. At the 
same time, effective mechanisms that link community and government must be operating in a 
conducive, enabling environment. It is this conceptual framework that underlies the comments 
below. 
Scaling up 
Interestingly, this study did not find that Mahavita suffered from many of the potential obstacles 
to scaling up, such as community or staff capacity, perhaps because Mahavita itself avoided 
many of them. Among the most important facilitating factors: 
• Mahavita developed a vision for community development that it promoted and, 
ultimately, had shared by key stakeholders. 
• Upper-level stakeholders, like CARE’s country director and USAID, provided space for 
the development of a community-driven approach (despite some serious misgivings and 
criticisms for not producing more outputs sooner). 
• Government policy supported decentralization, and few funds were otherwise available 
for fokontany development. 
• The project manager strongly supported the idea of community-driven development and 
pushed for the empowerment of his staff to mirror the empowerment he wished them to 
transfer to the community. 
• Mahavita developed processes and tools through pilots (and subsequent reflective 
learning) before expanding operations. 
Examination of some other issues in more depth reveals both strengths and weaknesses: 
Empowering Communities 
Members of the structure de developpement (SD) and the fokontany authorities were unanimous 
that Mahavita has benefited the community. A promising sign was that, when asked how they 
had benefited from Mahavita and the FAMOA process, by far the most common answer was the 
“change in mentality,” rather than the works as such. The FAMOA process had taught them how 
to work together as a community. 
Empowering Staff 
Mahavita has developed a learning culture, emphasizing learning-by-doing, reflection, and the 
positive aspects of learning from mistakes. This atmosphere empowers staff to be creative and 
take charge of their own activities. This sort of flexibility and empowerment, within the 
framework of principles, is essential for success in the field. Capable field staff who know 
community conditions are almost always better able than management to handle particular 
situations that arise at the community-level. 
Indeed, this kind of confidence in field staff, and a willingness to work closely with them to 
develop their abilities so they can exercise independent decision-making, is essential to a 
successful project. Centralizing decisions not only takes time as decisions work their way up and 
down the authority channel, without wide consultation and understanding this mode of operation 
also will likely not take adequate account of community conditions. In addition, staff respond to 
CARE Madagascar: Title II DAP Final Evaluation Page 52 
the management example they are given: If staff are disempowered, how can they be expected 
to work in an empowering way with communities? 
Establishing Processes 
Successful scaling up, and successful community-driven development, relies on developing 
principles for action, rather than a standardized model. Standard structures, mandates and 
mechanisms may be easily replicated but they violate the very spirit and benefits of development 
actually driven by the community. 
CARE did this in its development of a FAMOA process, which is extensively documented. All 
staff, and SDs and fokontany authorities, are well aware of this process, and can easily outline its 
steps. This facilitates scaling up because staff and communities now have a blueprint for action. 
Control of that action, however, remains in the hands of the community. 
The output of this community-driven planning process is the PDQ. FAMOA and the PDQ are 
obviously major successes, but Mahavita must now work to bring the process full circle. It must 
continue to evolve and adapt its approach based on its experience. 
As of now, although staff and SD do intuitively understand that communities should take the 
PDQ as a guide for action, the steps they should take post-PDQ are not clear. A number of 
communities are waiting to receive their “finalized” PDQ back from CARE; others have the 
PDQ and appear to see its use mostly to show other NGOs who happen to drop by. By default, 
then, the PDQ is an “end post,” and not a starting point, in the process of promoting community￾driven development. 
CARE and communities, however, must understand that the PDQ is part of a process. 
Ultimately it is a tool for community-driven development; it is not an objective. CARE must 
support the community in implementing its PDQ, probably through various iterations and 
exercises, but the goal is to enable communities with the capacity to carry out planning exercises, 
fund-raising, and implementation on their own. 
CARE must also work to affect the environment in which community-driven development can 
take place. This means advocating to raise awareness of government authorities of the structures 
de development and the existence of the PDQs as an initial output of these new community 
structures, and its usefulness in their own planning activities. As CARE did with the 
implementation, testing, and documentation of FAMOA, Mahavita should develop a strategy for 
supporting the priorities that result from the participatory planning process. Staff can employ a 
reflective process to experiment with approaches and document how to “keep the PDQ alive.” 
Supporting the Elements of CDD 
Even as Mahavita must develop its support of the SD and the PDQ further, Mahavita also needs 
to develop a strategy to strengthen the other components necessary for successful community￾driven development. Specifically, Mahavita now has to turn its attention to developing the 
capacity and mindset of government authorities to embrace and respond to community-driven 
needs and demands; and to the mechanisms that link community and governance structures. 
CARE Madagascar: Title II DAP Final Evaluation Page 53 
Although managers note that strengthening these components is a significant activity under the 
new DAP, more could have been done earlier. Many communities have had their final PDQs for 
a year or two. Although some CARE staff have worked to help the SDs find funding; and SDs 
have sometimes actively sought funds themselves, this is not a concerted, coordinated, or 
strategic effort to develop community capacity and keep the PDQ “alive.” Outcomes in the form 
of new projects or financing are limited. SDs certainly do not feel they have developed any 
insight into how to find funding; staff are not sure of how to interact with the SDs once the PDQ 
is done (many feel it is entirely up to the SD to work out how to implement the PDQ); and staff 
are not sure of where to turn for advice or information on funding opportunities within Mahavita. 
At the risk of putting staff and partners through yet another change, CARE should closely 
examine Mahavita’s organizational structure to ensure it supports these strategic thrusts. CARE 
must work on the “supply side” of government to support CDD. Government officials must 
strengthen their understanding of their responsibilities as public servants of all citizens and of 
their roles and their roles in a decentralized, democratic system (which remains to be clarified). 
Given limited financial and staff resources, government must also plan strategically to make the 
best use of them. Right now, CARE’s effort in this important area is weak. Only one person is 
responsible for institutional strengthening and advocacy beyond the community level. 
Formal and Informal Capacity Building 
Building capacity involves not only formal training but informally promoting knowledge 
exchange and strengthening institutional structures and administration. CARE’s strategy here is 
not clear to partners or staff. Staff were not sure whether “training” was independent or part of 
specific interventions; nor could they articulate content or purpose well. Some SD members and 
FKT authorities did mention training, but most were not sure whether the SD had received any 
training – and these community leaders should be key targets for institutional strengthening. In 
any case, training in issues like democracy and good governance as well as project management 
is clearly not a sustained or strategically targeted effort within Mahavita. 
Clarifying Roles of the SD and the FKT Authorities 
The institutional distinctions between the SDs and the FKT authority are not clear to staff and to 
stakeholders (although they can be clear about the different roles and responsibilities). In the 
communities themselves, the main distinction that has emerged is that the FKT authority is the 
“administrative” agency and the SD the “implementing” agency. But whether then SD responds 
directly to the FKT authority seems to vary by fokontany. In this exploratory exercise, we 
found that in fokontanies that staff said “worked best” the FKT authorities seemed to believe that 
the SD was essentially under their control (a concept reinforced by the fact that many of those 
interviewed held leadership positions in both). Those fokontanies that staff said “had problems” 
were those with political conflict between the two. Interviews with SD and FKT authority 
members verified that the two institutions saw themselves as being independent of one another. 
(Note that conflict could still arise in these fokontanies even if people held positions in both, 
because the “institutional conflict” could relate more to specific individuals.) 
Some confusion may have arisen during interviews because it was difficult to clarify exactly 
what was meant by the FKT authority “controlling” or “regulating” the SD (because, in a broad 
sense, any government authority would do exactly that over any private or non-governmental 
CARE Madagascar: Title II DAP Final Evaluation Page 54 
entity). But the significant overlap between the leaders of both and the reports of staff of 
political conflict confirms that this is a genuine issue, if not yet a problem. The exact 
relationships that currently and should exist need to be investigated further. 
To now, Mahavita has not tried to clarify the relationship between the SD and the FKT 
authorities for two main reasons. First, Mahavita likes to see “what develops,” and let 
communities go their own ways, instead of imposing guidelines. Second, trying to separate 
FKT authorities from the SDs may have created political problems for Mahavita right from the 
beginning. So Mahavita has actually encouraged FKT authorities to be part of the SD. 
Mahavita’s start-up strategy involved not only getting FKT authorities’ approval of the project 
but essentially naming the core FKT authorities as the initial ad hoc SD. Later, other influentials 
were added to the SD, but the core members began with (and often remain) the FKT authorities. 
Although this encourages initial FKT approval, in the longer term it runs a high risk that the SD 
effectively becomes simply another, probably inert, government agency and those political elites 
or other influentials will capture the SD for their own party or personal purposes. The focus 
groups did not look into the issue of whether this is already beginning to happen, or whether the 
SD is truly a voice of the community, and especially of the poor. This is clearly an important 
area for future research. 
The Scope of Responsibility of the SD for Community Planning
The relationship of the SD to the planning process and other NGOs is another issue that needs 
clarification. As noted, most actors currently see the SD as having the responsibility for 
planning and implementing community development. This implies that activities of other NGOs 
should also fall under its purview. And FKT authorities and SD members agreed that FKT 
authorities generally told any NGO that wanted to work in the area to first see the SD. 
CARE staff have been careful to stress that the SD is a creation of the community, and not of 
CARE. Still, CARE’s role in setting up the SD is undeniable. So far, however, other 
organizations do not seem to have objected to the SD’s assertion of this authority over them, 
even though it is entirely unofficial. 
This may be for a number of reasons: that this is a theoretical and not actual power and the SD 
makes no attempt to exert this authority (in fact, no one actually mentioned an instance where the 
SD was taking on responsibility beyond simply asking to be informed); no NGOs exist in the 
area, or they are not particularly strong; CARE staff work with other NGOs to help them 
“understand” why and how they should coordinate with the SD; the NGO has a community 
contact who is also a member of the SD and who supports central coordination; or the NGO 
ignores the SD. 
For now, the SD seems to be playing this role without objection, but that situation is likely to 
change once more resources are at stake. The SD actually has no official authority to plan or to 
supervise and coordinate the activities of others. The FKT authorities would. As the SD works 
to implement its PDQ, how these relationships with the FKT authorities and other development 
organizations in the community play out become critical to the successful implementation of the 
PDQ and to the very conception of what the SD is and from where it derives its power. For now 
its authority rests on (perhaps) perceived competence and filling an important need. Those are 
CARE Madagascar: Title II DAP Final Evaluation Page 55 
fragile bases. Mahavita needs to consider how to develop the SD institutionally or risk later 
disintegration of the SD. 
Formalization of the SD 
One potential model is to give the SD formal stature of some sort in the coordination and 
planning of community development. Legal statutes could recognize the SD as the CBO within 
the community that gives voice to residents during the government’s planning process (not that 
this would preclude other individuals or organizations from speaking up). But the government 
could formally require planners to give the SD’s point of view “great weight.” In addition, if the 
SD continues to produce high-quality PDQs, more and more government officials may see the 
advantages of working with the SD: to legitimize their actions as representing the community 
and to make their own work more efficient by building on the PDQ in making plans at higher 
area levels (the arrondissement, the commune). 
In fact, many SDs are already establishing their organizations officially. Although CARE staff 
are often encouraging and advising this process, Mahavita has no formal policy or guidelines 
regarding formalization. As of now, the process and Mahavita’s support of it is ad hoc. Staff are 
proceeding without any clear idea of what they are doing or how to do it. For such an important 
issue, CARE needs to consider the pros and cons of various institutional arrangements, assess the 
“experiments” that are already going on, and decide on an organizational strategy. 
Confusing Institutional Environment. 
The governance environment above the level of the fokontany further confuses the situation. 
Currently Madagascar has two parallel structures for planning at the municipal level (the Prefet 
and the mayor’s office), limited resources and confusion over who controls those resources, and 
no policy that outlines responsibilities for administrative structures below level of the 
municipality. For CDD to work effectively, the community must know who the key decision 
makers who affect their lives are. The PDQ will remain an effectively dead document despite 
the best community efforts if the SDs and FKT authorities do not know how it “plugs into” the 
planning process of upper-level authorities. More importantly, if the process is muddled, those 
in favor of more robust democratization and good governance will have a difficult time figuring 
out what regulations and which people are the key ones to target in advocacy efforts. Scaling up 
reliably and efficiently must take place where actors share understanding of a clear institutional 
framework and agree on institutional roles. The strategy, of course, must take into account the 
fact that SD members come and go, and that government authorities, from the FKT level up, 
change often. 
Expansion: Other Constraints and Possibilities 
Under the next DAP, CARE foresees expanding from the 30 fokontanies where it now works to 
80. But there are 192 fokontanies in Antananarivo. It also plans to expand to Fort Dauphin. 
Although this expansion will not happen overnight, and it can proceed with expansion to the 
other areas using the FAMOA approach in which it has confidence, CARE has not completed 
“the process” in the first 30’; it must do so to ensure it leaves a sustainable CDD process in 
place. As emphasized above, it needs to bring its support for CDD full-circle and work on the 
“supply side” and the institutional environment to ensure that all the elements of successful CDD 
CARE Madagascar: Title II DAP Final Evaluation Page 56 
are in place. In accomplishing its goal, CARE will have to test and document approaches 
(including an exit strategy), and build capacity among staff and stakeholders. 
CARE should also think about how to work with communities, other NGOs, and government to 
expand the reach of CDD within Antananarivo. Does CARE share its approach with other 
national and international NGOs? Can other approaches to CDD be contained within CARE’s or 
can they be synthesized? Does the government support the establishment of SDs in other 
fokontanies? Can SDs and fokontany authorities themselves build networks among themselves, 
where all share experiences and the strongest and most capable lead the establishment of CDD in 
other areas? 
CARE should build on its own knowledge and experience with these communities to maintain its 
focus on establishing and building the capacity of the SDs and shaping the enabling 
environment. This is CARE’s comparative advantage. This should not be the responsibility of 
other international or local NGOs. Their intervention as a “replicator” of FAMOA would largely 
only add another layer of organizations among which CARE would need to build capacity, as 
many NGOs have program-specific expertise (such as housing or health). Rather, they should be 
seen for what they are: additional sources of financing and program expertise (usually revolving 
around very specific activities such as housing or health) that the SD can tap into to realize its 
own PDQ. CARE can also continue to build capacity among the SDs, who can themselves, 
rather than NGOs, replicate the model through sharing their experiences with other FKTs. 
As CARE scales up, it must seriously consider the impact on staff. The numbers in the DAP are 
certainly guides, but management must be careful not to expect too much too soon. This is not a 
lowering of expectations, but a realization that working with SDs and other stakeholders to 
establish common visions, institute structures, and resolve conflict is not easy and requires a 
significant investment of time. Supporting true institutional development of the other 
stakeholders and changing the institutional environment are activities new to CARE. CARE 
must focus on going slowly and leaning as it goes, to ensure sustainability. Management should 
carefully monitor staff to see how things go and ensure sustainable structures and processes are 
in place before moving on. 
Staff, SDs, and FKT authorities consistently mentioned that one of the most difficult tasks in 
promoting the FAMOA process was that community residents wanted immediate actions and 
works. Such changes take time, and now it will not only try to do this in an additional 50 
communities but will take on the task of changing the mentalities of authorities in the 30 where it 
has already worked. If CARE simply sees the next DAP as expanding the current approach, with 
its focus on the PDQ as its output, and does not spend time to “complete the circle,” it will fail in 
its efforts to establish sustained community-driven development in any of the 80 FKTs. 
Part of moving in this way is establishing the appropriate benchmarks for CDD in its role as a 
facilitator. Certainly CARE will have to respond to the standard operating indicators of donors 
(although over time I believe a coalition of donors, research organizations, and development 
agencies, including NGOs, will move donors like USAID to modify these so that, for facilitating 
organizations, they focus more on outputs related to process than to global development 
indicators). But this should not limit CARE’s efforts to developing its own indicators that it 
CARE Madagascar: Title II DAP Final Evaluation Page 57 
needs for its own management purposes anyway. CARE has apparently already moved 
somewhat down this path in its next DAP. CARE can continue to innovate in this area. And 
again, it makes sense just from its own internal management perspective to have appropriate 
indicators to measure its performance and the performance of other stakeholders and partners. 
A significant challenge will be ensuring that Mahavita maintains the rhythm of developing and 
solidly establishing CDD. CARE is leading this effort. Much of its success depends on being 
seen as a credible partner by the various stakeholders. If it fails to outline a process and keep its 
commitments, SDs, government authorities, and residents will lose interest and drift away. 
CARE has yet to clearly outline its final goal (after the exit) for communities and authorities, and 
share its strategy to achieve that (strengthening all the elements). Although all communities feel 
they have developed a sense of solidarity and confidence in their ability to act through the 
FAMOA process, if CARE does not lead the further development of CDD, it risks losing that 
goodwill and the building blocks, already constructed within the community, for CDD. 
Some incidents already raise flags: CARE’s limited support for building institutional capacity 
and designing strategies to implement the PDQ that leaves the communities rather stuck; some 
mention among SDs and FKT authorities that in carrying out projects CARE technicians made 
decisions without consulting the community leaders or explaining their reasoning to them; and 
others who commented that sometimes, deliveries of construction inputs were late or never 
arrived. 
Building community relations is essential to the work of establishing CDD, but it is a fragile 
process. Experiences elsewhere show that building and maintaining credibility among these 
vulnerable communities is key. They have had hopes dashed when working with outsiders and 
experienced the disdain of those who are wealthier and more powerful. CARE promises 
something different. For CARE to function effectively as a change agent, it must keep that 
promise. 
Scaling Up: A Checklist 
Experiences of scaling-up highlight a number of key issues that affect the nature and success of 
attempts to scale up (Gillespie 2002). This review of Mahavita points out the importance of 
some of these issues. Community-driven development often happens as a result of some outside 
catalyst. Here the need to develop a DAP was a trigger for development of Mahavita. Upon her 
arrival, Lisa Dean, the CARE country director, found a fairly dysfunctional country office with a 
crying need for funds. 
Individuals in authority with vision are often key to taking advantage of this opportunity. This 
was the case with Mahavita as well. Michael Drinkwater, CARE’s regional adviser, transferred 
community-driven experiences from Zambia to Madagascar. Chris Dunston joined Mahavita 
and continued the focus on community and staff empowerment, emphasizing the importance of 
establishing processes, of flexibility in implementation, and of learning. 
The political, institutional, cultural, and institutional environment affects the ability to which 
these individuals can realize their vision. Here the government was relinquishing some control 
(a legacy of a socialist past) and emphasizing decentralization and community participation. 
With significant education efforts on the part of CARE, USAID was permitting the project to go 
CARE Madagascar: Title II DAP Final Evaluation Page 58 
forward, albeit often complaining about the need for more concrete outputs. The environment 
thus gave space to the efforts that CARE was making. Changes in that environment can limit or 
promote community-driven development, so CARE will need to continue to pay attention to the 
legal and institutional challenges now found in Madagascar. 
The capacity of stakeholders is a principal determinant of success. Somewhat delayed, Mahavita 
is just beginning to confront this constraint. CARE staff themselves need to have a more 
complete idea of their actions in establishing community-driven development, beyond the PDQ. 
The community structures must also be strengthened, in concrete and philosophical ways. As 
many of those interviewed mentioned, changing the mentality of community residents to accept 
responsibility to be major players in their own development is a major undertaking, and as 
essential to success as ensuring the structure de development has the technical, administrative, 
and financial capabilities to function effectively. 
The building block of scaling up, however, is the development of empowering processes that can 
be used to guide expansion. CARE has made a good start with the development of FAMOA, but 
it must complete the circle. Community-driven development is the objective, not the PDQ. And 
in this context, key questions remain. How will CARE really operationalize the SD, ensure it 
represents the voice of the community and link it to outside resources and structures? What will 
it do to develop capacities and shape a stable, enabling environment? The key constraint on 
scaling up seems to be that CARE still needs to develop and test what processes and strategies 
will most successfully do these things. 
References
Dongier, P., J. van Domelen, E. Ostrom, A. Ryan, W. Wakeman, A. Bebbington, S. Alkire, T. 
Esmail, and M. Polski. 2002. Community-Driven Development. World Bank. Mimeo. 
Gillespie, S. Scaling Up Community Driven Development: An Overview. 2002. IFPRI. 
Mimeo. 
Uvin, P. 1995. Fighting Hunger at the Grassroots: Paths to Scaling Up. World Development 23 
(6): 927-939 
CARE Madagascar: Title II DAP Final Evaluation Page 59 
ANNEX B 
Evaluation Plan / Scope of Work 
CARE Madagascar’s Title II Development Assistance Program (DAP) 
1. Background and Purpose of the Evaluation 
CARE Madagascar’s DAP 1998-2002 – Transfer Authorization (TA) Award No. FFP-G-00-98-
00060-00 – was granted a ‘no-cost extension’, through the approved FY 2000 DAP Amendment, 
whereby advancing the official start-date by one year, to FY 1999 (instead of FY 1998), 
subsequently extending the life-of-activity (LOA) to the end of FY 2003. CARE Madagascar 
established two programs out of the DAP: Mahavita and CYPREP. Program Mahavita (Urban 
Household Livelihood Security) is addressing Strategic Objectives 1 and 2. Program CYPREP 
(Emergency Preparedness) is addressing the Special Program Objective. 
Program Mahavita 
Program Mahavita’s actions address food security in urban areas of Antananarivo by sustainably 
increasing the access of beneficiaries (poor urban households) to food, and by sustainably 
improving the utilization of food by beneficiaries. 
Improved access to food is obtained by offering income generation opportunities to project 
beneficiaries. This will allow them to purchase more food, and to build up assets. Asset buildup 
will help households overcome periods of difficulties. 
Improved utilization of food is fostered by protecting people’s health, via the construction of 
sanitation infrastructure, and the training of beneficiaries in improved sanitation, health and 
nutrition practices. 
When the DAP began implementation (October 1998), TOUCH 2000, a USAID funded child 
survival project had just completed its mid-term evaluation. The evaluation team strongly 
recommended that the two projects integrate activities and build in the results already achieved 
by TOUCH’s 2 years of implementation. TOUCH was successful in creating strong community 
support and motivation towards child survival initiatives. It was clear that the DAP could benefit 
from TOUCH’s efforts and lessons learned, particularly in terms of community mobilization and 
institutional capacity building methods. For example, most of the CACs had been very active, 
planning and carrying out health promotion activities. Also, through a number of participatory 
workshops, TOUCH was successful in creating a strong partnership with the Ministry of Health, 
particularly with those supervising and staffing health clinics. On the other hand, TOUCH 
initiatives would also benefit from DAP’s efforts in environmental health and personal 
empowerment objectives. Hence, the two projects were combined. 
CARE Madagascar: Title II DAP Final Evaluation Page 60 
2. Program Overview 
Mahavita’s baseline survey was conducted in March/April 2000. This survey included all the 
TOUCH results indicators. The final evaluation will include TOUCH activities as well and is 
planned to begin in February 2003 and conclude in September, involving 3 components. 
Two strategic objectives, which pertain to Mahavita activities, will be evaluated. There are as 
follows: 
Strategic Objective 1: Protect and promote the food and livelihood security of 10,000 urban 
households. 
SO 1 has two sub-objectives: 1) Income Generating Activity, and 2) Institutional Capacity 
Building. Result (impact, effect and monitoring indicators are as follows: 
1) Income Generating Activity 
• Index of key assets owned; 
• Average number of days a HH worked over previous two weeks; 
• Number of HHs who participated in FFW programs; 
• Number of HHs who received Personal Empowerment Training (PET). 
2) Institutional Capacity Building 
• Number of Structure du Development' (SD) operating sustainably; 
• Proportion of infrastructure maintenance costs covered by public; 
• Number of paid work-days generated by SD activities; 
• Number of SD established; 
• Number of SD with a Community Development Plan (CDP) in place; 
• Number of SD that have negotiated projects; 
• Number of SD that have completed projects. 
Strategic Objective 2: Sustainable improvement in environmental health in 30 Fokontany for 
200,000 people in urban Antananarivo. 
SO 2 has two sub-objectives: 1) Sanitation Infrastructures, and 2) Health and Nutrition 
Practices. Result (impact, effect and monitoring) indicators are as follows: 
1) Sanitation Infrastructures 
• Improved communication within fokontany; 
• Proportion of canals rehabilitated; 
• Number of HHs affected annually by floods; 
• Percentage of HHs using safe water storage container; 
• Number of people per water point; 
CARE Madagascar: Title II DAP Final Evaluation Page 61 
• Number of HHs per adequate garbage collection facilities; 
• Number of HHs per sanitation (toilets) and hygiene (bathing) facilities; 
• Percentage of HHs using latrines for human waist; 
• Meters of canal rehabilitated; 
• Number of public water points established; 
• Number of public sanitation (toilets) and hygiene (bathing) facilities established; 
• Number of garbage collection points established; 
• Number of washing basins established. 
2) Health and Nutrition Practices 
• Percentage of population experiencing diarrhea 2 weeks prior to survey; 
• Household dietary diversity; 
• IEC sessions on proper infant feeding practices; 
• Safe Water System Campaigns; 
• Number of HHs adopting Safe Water System Practices. 
From January – August 2002 Madagascar experienced a political crisis due to elections. During 
this time road blocks were common and fuel and food could not reach the capital Antananarivo. 
This resulted in increases of food and loss of jobs. Due to the capacity which existed throughout 
the operational zones of Mahavita, we were able to double our infrastructure efforts and 
programmed significant more food and cash through safe net programs. 
The objectives of CYPREP were designed as a Special Project Objective: Safeguard the 
livelihood security of Madagascar's most cyclone-vulnerable communities by enhancing local 
capacity to plan for and cope with emergencies. This objective involves two sub-objectives: 
• Number of villages that have a 'Disaster Preparedness Action Plan' in place – In other 
words: A 'Disaster Preparedness Action Plan' IN PLACE is defined as a village who has 
completed the planning process, has a formal 'Disaster Preparedness Committee', a 
'Disaster Preparedness Action Plan' and implemented at least one mitigation project e.g. 
school, grainery, socially marketed products mobilization and sales campaign. 
• Number of villages that have developed a 'Disaster Preparedness Action Plan' – In other 
words: A 'Disaster Preparedness Action Plan' is defined as a village that has completed 
the planning process with an informal committee. 
It was not necessary to conduct a baseline survey for CYPREP, due to the nature of the activities 
and the fact that the committees and mitigation strategies didn’t exist before. Due to the crisis, 
CARE had to decrease some of its activities do to the political problems throughout Tamatave 
and the lack of fuel. 
3. Analytical focus of the evaluation 
Mahavita's community-drive development approach was designed towards building institutions 
capacity to diagnose their problems and plan to solve them. The sustainability of the strategies is 
CARE Madagascar: Title II DAP Final Evaluation Page 62 
sought by organizing community development structures that generate employment around the 
satisfaction of public needs (namely, the construction and maintenance of public sanitation 
infrastructures). It is expected that the public’s satisfaction with the services provided will create 
a willingness to pay for the continuation of those services, thus maintaining a demand for the 
jobs created by the project, and a means to pay for those jobs. The objectives of offering income 
generation opportunities and of protecting health are thus closely related, and are mutually 
supportive. Questions: 
1. What has been Mahavita’s experience so far in expanding this approach to other 
communities while phasing out its assistance to the initial communities? 
2. What have been the key challenges and elements of success, especially in terms of 
building community capacity, addressing constraints in the institutional environment, 
taking an intensive pilot project to larger scale, and developing a strategy around 
processes and structures rather than projects? 
3. In the end, what do these experiences suggest that CARE and like organizations should 
do to encourage and expand sustainable community-driven development? 
4. What is the community members (households, and municipality) level of satisfaction of 
services performed by the SDs and their willingness to pay for their services. 
The household livelihood security baseline survey, conducted in March/April 2000, resulted in a 
quasi-model of urban poverty. Data collected about household asset ownership resulted in the 
calculation of a new variable representing total household assets. The data supports the 
hypothesis of asset accumulation serving as a proxy for household wealth. The asset ranking 
correlates well with variables such as living space (in square meters), number of rooms in the 
house, average expenditures for food and medicine, and education level. 
There was surprisingly less correlation with other variables, in particular those associated with 
health and hygiene behavior, number of days absent from work, and vaccination of children. 
There was evidence that suggests health behaviors are similar, though not exactly the same, 
across asset classes. There is also some health evidence to confirm the unhealthy environments 
cut across social classes, such as the percentage of each class reporting at least one case of 
cold/flu in the household. 
As a result, the survey identified several implications for urban programming which should be 
addressed in the evaluation; what efforts have been made to learn more and validate these 
programming implications and how activities have effected change. They are as follows: 
1. The project should explore opportunities for keeping kids, especially adolescents, in school 
in order to improve their future chances of gainful employment. 
 
2. The project should conduct focus group interviews to gain more in-depth knowledge about 
what drives diet diversity. The focus group discussions can target fruit, egg and dairy 
consumption to determine the consumption dynamics of these food items. The role of street 
foods in family nutrition should also be explored. 
CARE Madagascar: Title II DAP Final Evaluation Page 63 
3. The project should explore the reason behind high food expenditures. The high 
consumption of street foods may be a contributing factor, but little is known at present about 
the relative costs of eating at the home or outside of the home. Little is also known about 
the nutritional costs and benefits of urban Antananarivo eating habits. 
4. Cash for Work (or a mix of Cash for Work/Food for Work) programs should target the 
unemployed (in this survey 7% of all households) and those occupied in petty trade who live 
day to day in hopes of selling articles on the street. It should also take into consideration 
those Fokotanys where unemployment may be highest. 
5. According to survey results, 30% of all households in the project area are headed by a 
woman. Analysis also suggests that these households are among the poorest. Project 
interventions that target and benefit these households should be considered. 
6. Respiratory illness is high among all asset groups, suggesting that the urban environment 
does not discriminate among wealth rankings with respect to this health concern. However, 
households do vary in their behavioral responses to illness, with better-off households 
spending more money on health care and accessing better health services than other 
households. 
The environmental health component involves social and technical activities. The evaluation 
will need to address the following issues: 
• Engineering aspects (pre-post surveys of canal, walkway rehabilitation) 
• Materials and equipment management 
• Food management 
• Ration size and contribution to household nutrition 
• Life skills training courses 
• Income generating activities 
4. Proposed Evaluation Methodologies 
The evaluation will start with the qualitative assessment of the community driven development 
approach. This will be in the form of focus group discussions, individual interviews and result in 
a case study as well as a document providing a description of each issue and provide 
recommendations. 
The second component of the evaluation will be the post-test of the quantitative household 
survey conducted in March/April 2000. The sampling frame must be reviewed and designed so 
that the results will be able to be disaggregates by Fokontany. It will be important to look at the 
‘level of effort’ invested by each Fokontany and compare that with the changes in household 
livelihood security. The questionnaire used before will be reviewed and edited, making 
adjustments according to the level of significance in baseline monitoring results. The survey will 
be conducted by a local consulting firm, providing the final data set and preliminary analysis 
(cleaned data set, completed questionnaires, means, cross tabs, etc.). TANGO International will 
be responsible for doing the final analysis and report writing. 
CARE Madagascar: Title II DAP Final Evaluation Page 64 
The third component will be to evaluate the Environmental Health activity that involves 
cash/food for work to rehabilitate canal and walkways, and water provisioning through cost￾recovery systems. This will mainly be a tabulation of monitoring reports to determine the final 
numbers of food for work participants (man-days), kilometers of canal, walkways, etc., number 
of trainings, employment days generated, income earned and capacity of SD’s to manage/sustain 
these services. This will also involve a thorough review of the systems and procedures used in 
planning, implementation and monitoring activities related to civil engineering, materials and 
equipment and food management. The effectiveness of the cost-recovery systems will also be 
evaluating. This will be primarily quantitative; however, it will also include qualitative methods 
to evaluate the organizational and management aspects of the cost-recovery systems. 
To determine the effects of the Special Program Objective (CYPREP) an external evaluation 
team will be recruited to determine whether the results have been achieved. This will primarily 
involve a quantitative review of the anticipated outputs stated in the original agreement. It will 
also involve a qualitative sample of Community Cyclone Preparedness Committees to determine 
the ‘state-of-affairs’ of the plans and outcomes related to mitigating activities. 
 
CARE Madagascar: Title II DAP Final Evaluation Page 65 
ANNEX C 
Persons Contacted 
FRIEDEBERG Tom CARE/Madagascar Country Director 
DANIEL Jean DAP Coordinator 
RAMANARIVO Leandre Chef de Projet Mahavita 
ANDRIAMISANDRATRA Jules Chef de Departement, Institutional Capacity 
Building 
Dr. RALAMBOSON Bodosoarivelo Coordinatrice Sante, Projet Mahavita 
HOLIHASINORO Sabine Coordonnatrice de Zone/Responsible Renforcement 
de Capacite 
RATOARIJOANA Avo Assistante IEC 
RAHERIJAONA Hanitra ADC 
Joelle ADC Ambilanibe 
ANDRIAMAHEFARIVO Landy Responsable des Infrastructures Generatrices de 
Revenue 
RAKOTOARISOA Naina Responsible des Infrastructures 
Francia Assistante de Volet Infrastructure 
RAMANANDRIMALALA Rija Agent de Terrain, Mandrangobato II 
RAVELOMANANTSOA Solofo ADC Ankasina, Antaniavo, Andrahavo 
RABEKOTROKO Fidy ADC Tsaramasay 
RAKOTOLANDRY Lalaine Coordonnateur de Zone/Responsible Renforcement 
de Capacite 
Members Structure de Developpement: Tsaramasay 
 Andavamamba 
 Antetezana Centre I 
 Ankasina 
Dr. RAMANANDRAIBE Lucie Chef de Service, Direction de Soins Medicaux et 
d’Appui aux Districts Sanitaires 
Dr. RAVAOMANANA Anne Marie Inspecteur de Sante, Antananarivo 
Dr. RAKOTOARIMANANA Hanitsoa Medecin Chef, CSB de Isotry 
Dr. RAHELIMALALA Robertine JSI-RTI 
Wendy BENAZERGA USAID/HPN 
Pedro CARILLO USAID Food Security and Disaster Manager 
Chris DUNSTON former Mahavita DAP Coordinator 
 
CARE Madagascar: Title II DAP Final Evaluation Page 66 
ANNEX D 
List of Antananarivo Fokotanys Assisted by CARE’s Title II Program 
Arrondissement I:
Ankasina 
Antohomadinika FAAMI 
Antohomadinika Nord Antaniavo 
Antohomadinika III G Hangar 
Antohomadinika Centre 
Antohomadinika Sud 
Antetezanafovoany I 
Antetezanafovoany II 
Manarintsoa Anatihazo 
Ambalavao Isotry 
Andavamamba Anatihazo I 
Andavamamba Anatihazo II 
Andavamamba Anjezika I 
Andavamamba Anjezika II 
Andohatapenaka I 
Arrondissement III:
Ankazomanga Nord 
Tsaramasay 
Ankorondrano Ouest 
Ankorondrano Andranomahery 
Arrondissement IV:
Ivolaniray 
Andavamamba Ambilanibe 
Anosibe Ouest II 
Anosibe Mandrangobato I 
Anosibe Mandrangobato II 
Anosibe Angarangarana 
Madera Namontana Est 
Anosibe Ambohibarikely 
Ouest Mananjara 
Arrondissement VI:
Andraharo 
Ankazomanga Sud 
CARE Madagascar: Title II DAP Final Evaluation Page 67 
 
ANNEX E 
Sample Sites from the Qualitative Survey - CYPREP 
Zone Sous-Préfecture Commune Fokontany 
Nombre de 
types 
d’activités 
de 
mitigation 
Activités de 
mitigation 
Tamatave II Tamatave 
Suburbaine 
Ranomena 1 Abri anti-cyclone 
Brickaville Ambodifaho 2 - Aménagement 
- Adduction d'eau 
-
Andovoranto Andovoranto 0 -
TAMATAVE 
Brickaville 
Mahatsara Vohiboazo 1 - Pistes 
Mahanoro Sahabe 2 - Aménagement 
- Pistes 
Tsaravinany Taviranambo 2 - Abri anti-cyclone 
- Pistes 
Ambodiharina Maromitety 1 - Adduction d'eau 
MAHANORO Mahanoro 
Masomeloka Valokianja 0 -
Fénérive-Est Ambodimanga II Ampasimpotsy 2 - Reboisement 
- Pistes 
Marofinaritra 2 -Grenier 
- Pistes 
Vavatenina 
Mahatera 2 - Adduction d'eau 
- Pistes 
Vavatenina 
Anjahambe Anjahambe 4 - 2 abris anti-cyclone
- 2 greniers 
-reboisement 
- Pistes 
Tamatave II Foulpointe Marofaria 0 -
FENERIVE￾EST 
Soanierana 
Ivongo 
Soanierana 
Ivongo 
Tanambao 
Matsokely 
1 - Abri anti-cyclone 
CARE Madagascar: Title II DAP Final Evaluation Page 68 
ANNEX F 
Details of Qualitative Research Visits 
Zone Tamatave 
Date Lieu Personnes rencontrées Observations
Lundi 18 août matinée Tamatave II Sous-Préfet Absent 
 Tamatave Sub-Urbaine Maire 
 Bureau CARE Tamatave AMS 
 après-midi Ranomena CVS 
 Bénéficiaires 
Mardi 19 août matinée Brickaville Adjoint du Sous-Préfet 
 Maire et Adjoint au Maire 
 Mahatsara Maire et Adjoint au Maire 
 après-midi Vohiboazo CVS 
 Bénéficiaires 
Mercredi 20 août matinée Andevoranto Adjoint au Maire 
 CVS 
 Bénéficiaires 
 après-midi Ambodifaho CVS 
 Bénéficiaires 
Jeudi 21 août matinée Bureau CARE Superviseur Tamatave 
Zone Mahanoro 
Lundi 25 août matinée Départ Mahanoro 
 Mahanoro Sous-Préfet et Adjoint au Sous-préfet 
 après-midi Sahabe CVS 
 Bénéficiaires 
Mardi 26 août matinée Tsaravinany Maire 
 Taviranambo CVS 
 Bénéficiaires 
 après-midi Mahanoro Maire 
 AMS 
Mercredi 27 août matinée Ambodiarina Maire 
 Maromitety CVS 
 Bénéficiaires 
 après-midi Trajet Masomeloka 
Jeudi 28 août matinée Valokianja CVS 
 Bénéficiaires 
 après-midi Retour Mahanoro 
 Bureau CARE Superviseur Mahanoro 
 Bureau CARE Maire de Masomeloka 
Vendredi 29 août matinée Bureau CARE Mahanoro Coordinateur Programme Mahanoro 
 Administration 
 après-midi 
Samedi 30 août matinée Bureau CARE Mahanoro Restitution 
 après-midi Retour Tamatave 
CARE Madagascar: Title II DAP Final Evaluation Page 69 
Zone Fenerive Est 
Date Lieu Personnes rencontrées Observations
Lundi 01 sept matinée Foulpointe Maire 
 Marofaria CVS 
 Bénéficiaires 
 après-midi Fénérive Sous-Préfet Absent 
 Ambodimanga II Maire 
 Fénérive AMS 
Mardi 02 sept matinée Vavatenina Deux Adjoints du Sous-Préfet 
 Maire 
 Anjahambe Maire 
 après-midi Anjahambe CVS 
 Bénéficiaires 
Mercredi 03 sept matinée Marofinaritra CVS 
 Bénéficiaires 
 après-midi Mahatera CVS 
 Bénéficiaires 
Jeudi 04 sept matinée Tanambao Matsokely CVS 
 Bénéficiaires 
 après-midi Soanierana Ivongo Sous-Préfet 
 Adjoint au Maire 
Vendredi 05 sept matinée Ampasimpotsy CVS 
 Bénéficiaires 
 après-midi Bureau CARE Fénérive Tovo : Ancien superviseur de Tamatave 
 Superviseur Fénérive 
Samedi 06 sept matinée Bureau CARE Fénérive Restitution 
 après-midi Retour Tamatave 
Dimanche 07 sept après-midi Tamatave 
Présentation du déroulement des travaux 
à Richard Caldwell 
Lundi 08 sept matinée Tamatave 
Réunion avec l'équipe de recherche 
quantitative 
Lundi 15 sept journée Tamatave Interview du Chef de projet 
Mardi 16 sept matinée Tamatave Administration Tamatave 
 après-midi Interview Coordinatrice du Programme Est 
CARE Madagascar: Title II DAP Final Evaluation Page 70 
ANNEX G 
CARE / PROGRAMME MAHAVITA
EVALUATION DES TRAVAUX D ' INFRASTRUCTURES SOCIALES REALISES 
SELON LE PROGRAMME MAHAVITA 
A. INTRODUCTION 
- Financement : USAID 
- Gestion du projet : CARE INTERNATIONAL / PROGRAMME MAHAVITA 
- Objectifs : Evaluation des travaux d'infrastructures sociales réalisés selon le programme 
MAHAVITA 
- Zone d'intervention : Quartiers bas où la majorité de populations est pauvre 
- Mission : Apprécier le niveau d'atteinte des objectifs relatifs à l'amélioration des 
infrastructures communautaires 
- Evaluer les capacités techniques des partenaires Mahavita 
- Evaluer l'organisation de l'unité de programme Mahavita en charge des aspect 
d'engineering 
- Recommandations nécessaires pour la suite du programme 
 Demarche de l'Evaluation 
- Un programme de visite du 23/09/03 a été réalisé avec les Responsables CARE sur les 
Fokontany suivants: 
Fokontany Infrastructures
 Mandrangobato I Ruelles 
 Mandrangobato II Diguettes 
 Anosibe Ouest II Borne fontaine et bloc sanitaire 
 Ambilanibe Bassin lavoir 
 Anatihazo II Bassin de rétention 
- Evaluer les ouvrages réalisés suivant les plans fournis par le Responsable des 
Infrastructures 
- Les programmes des travaux réalisés par les communautés sont : 
 1°) Bassin de retention 
 2°) Aménagement diguettes 
 3°) Ruelles en béton 
 4°) Passerelle en bois 
B. NIVEAU ATTEINT SUR L'AMELIORATION DES INFRASTRUCTURES 
COMMUNAUTAIRES 
B1. Bassin de Retention
- Il est à signaler que les bassins sont réalisés en collaboration avec L ' APIPA et les 
techniciens du Génie Rural 
CARE Madagascar: Title II DAP Final Evaluation Page 71 
- Prestation de l 'APIPA avec les communautés : 
- Fourniture de camion pour le Transport de vivres et évacuation des produits de curage 
- Fourniture de pompe
Bassin de Retention a Anatihazo II
 
- Superficie du bassin réalisée : 60 000 m2 
- Objectif des travaux : Stockage des eaux pluviales pour lutter contre l ' inondation 
- Nombre de fokontany bénéficiaires : 15 dont 
- Anatihazo I et II 
- Ambilanibe 
- Anjezika I et II 
- Antohamadinika Sud 
- Manarintsoa Anatihazo 
- Ivolaniray 
- Anosibe Ouest II 
- Ambohibarykely 
- Angarangarana 
- Mandrangobato I et II 
- Madera 
- Ouest Manjara 
Qualité des travaux : 
- Profondeur du bassin : dans le norme 
- Evacuation des produits de curage : Une partie non réalisée (à la charge de l'APIPA) 
- Absence des panneaux pour les travaux à réaliser et de signalisation 
Diguettes côté Nord du bassin du de rétention: 
- Matériaux utilisés : Produits de curage en tourbe vaseuse plus ou moins argileuse 
- Compactage : quasiment mal compacté permettant le passage des eaux 
pluviales à travers la diguette à cause des retraits de l ' argile. 
- Absence de mottes de gazons sur la pente du talus 
- Pente des talus non respectée 
- Quelques buses non encore placées sur leur emplacement prévu 
- Couche de revêtement instable dans l'ensemble 
Recommandation
1. L'instabilité de la diguette peut provoquer pendant la période de crue de problèmes sur le 
bassin ainsi , pour éviter l'émergence de brèche au niveau de la diguette et l'instabilité du talus , il 
faut implanter au niveau de chaque pente de talus des mottes de gazons 
 
2. Mettre les buses avant la période de pluie sur les emplacements prévus 
3. Bien suivre le plan de maintenance 
CARE Madagascar: Title II DAP Final Evaluation Page 72 
Dans un prochain avenir : 
1. Apport des matériaux convenables pour avoir un bon compactage (remblai Latéritique). 
2. Pente à respecter pour assurer la stabilité. 
 
3. Engazonnement par mottes de gazons avec fixation à l'aide de piquets. 
 
4. Bien maîtriser les ouvrages d'équilibre. 
 
5. Evacuation des déblais de curage. 
 
6. Bourdure du bassin planté en vetiver. 
B.2 Evaluation Quantitative et Estimative de Bassin de Retention
 Superficie du bassin (m2) : 60 000 m2 
 Quantité de déblai à enlever 30 000 m3 
 Délai d'exécution 50 j 
 Nombre de MO 2600 
 Rendement 0,23 m3/j 
Quantite Prevue et Reellement Consommee
DESIGNATION UNITE QUANTITE 
PREVUE
QUANTITE 
CONSOM.
Nombre MO u 2 600 1 305 
Total Hj Hj 130 000 107 000 
Nombre Quartier u 7 15 
Outillages 
Panneaux u 4 0 
Petits outillages MO 2600 2600 
Matériaux 
Buse D300 u 28 28 
Ciment Sac 24 24 
Sable M3 10 10 
CARE Madagascar: Title II DAP Final Evaluation Page 73 
Cout des Travaux d'Amenegement du Bassin de Retention
DESIGNATION UNITE PU QUANTIT
E CONS.
MONTANT
Total Hj Hj 11000 107 000 1 177 000 
000 
Outillages 
Panneaux u 250000 0 0 
Petits outillages MO 10000 2600 26 000 000 
Matériaux 
Buse D300 u 75000 28 2 100 000 
Ciment Sac 60000 24 1 440 000 
Sable M3 30000 10 300 000 
TOTAL 1 206 840 
000