Final evaluation of the project Kore L’Avni Nou
Food Voucher Programme implemented by CARE in the department of Grande-Anse
By
Isnel Pierreval
Luckny Zéphyr
James Lachaud
(Diagnostic & Development Group/www.ddghaiti.com)
Submitted to
CARE HAITI
February 2013
2
Table of contents
1. Executive summary......................................................................................................................4
1.1 Beneficiary outcomes .........................................................................................................4
1.2 Program effectiveness.........................................................................................................5
1.3 Recommendations...............................................................................................................6
1.3.1 Main recommandations ............................................................................................6
1.3.2 Secondary recommendations....................................................................................7
2. Introduction..................................................................................................................................8
2.1 Background.........................................................................................................................8
2.2 Methodology.......................................................................................................................9
2.2.1 Quantitative methodology ........................................................................................9
2.2.2 Qualitative methodology .......................................................................................10
2.2.3 Sample size and data collection..............................................................................10
2.3 Limits of the evaluation ....................................................................................................12
2.4 Organization of the report.................................................................................................13
3. Evaluation results.......................................................................................................................13
3.1 Impact evaluation and comparison with the baseline study .............................................13
3.1.1 Socio-Demographic profile of the respondents......................................................14
3.1.2 Impact on beneficiary for food security .................................................................15
3.1.3 Effect on beneficiaries coping strategies................................................................16
3.1.4 Benefits on households assets ................................................................................19
3.1.5 Impact on beneficiary’s personal finances (debt and borrowing) ..........................22
3.1.6 Impact on education on beneficiaries' children education......................................23
3.1.7 Access to mobile phone..........................................................................................24
3.1.9 Effect on shops and on local economy ...................................................................24
3.1.10 Impact on shops' managerial capacity ..................................................................25
3.1.11 Impact on amount of money in local economy ....................................................25
3.1.12 Effect on local production and employment ........................................................26
3.1.13 Effect on shop coping strategies...........................................................................27
3.2 Overview of data collected on regular basis and presentation of performance monitoring28
3.2.1 Monitoring and evaluation system .........................................................................28
3.2.2 Data collected by the monitoring and evaluation staff...........................................30
3.3 Relevance and effectiveness of implementation of the electronic food voucher program
in Grand’Anse .....................................................................................................................32
3.3.1 Food insecurity in Haiti and Food distribution challenges.....................................33
3.3.2 Description of CARE electronic food voucher program........................................33
3.3.4 Effectiveness and relevance of beneficiary’s selection process.............................34
3.3.5 Relevance and Effectiveness of registration mechanism .......................................36
3
3.3.6 Relevance and Effectiveness of redemption (food delivery) .................................38
3.3.7 Preferred Form of food assistance ..........................................................................40
3.3.8 Satisfaction with the project by beneficiaries.........................................................41
3.3.9 Satisfaction with the project by shops....................................................................42
3.3.10 Program partners and sustainability .....................................................................42
3.3.11 SWOT analysis of the Food Electronic Voucher and scaling up as implemented45
3.4 Key lessons learnt .............................................................................................................47
3.5 Key obstacles preventing the correct execution of the program.......................................47
3.6.1 Challenges encountered by shops during the project .............................................47
3.6.2 Issues with the beneficiaries...................................................................................48
3.6.3 Issues with the community .....................................................................................48
3.6.4 Issues with shops....................................................................................................49
3.6.5 Issues with the technology......................................................................................49
3.6 Main recommendations ....................................................................................................49
4. Conclusion .................................................................................................................................50
4
1. Executive summary
This report presents the final evaluation findings on the “Food Voucher Programme - Kore
L’Avni Nou” implemented by CARE in nine (9) communes of the "Grande d'Anse" department in
Haiti. The program aimed to support 12,000 vulnerable families with a $50 (2,000 HTG)
monthly voucher over a six-month period. In addition, 68 local shops were approved to
supply the beneficiaries with ke y commodities: Rice, vegetable oil, maize and beans.
Beneficiaries were also able to purchase magi, haring, and fish within 10% of total cost of the
voucher. Merchants received payment through an electronic voucher platform developed by CARE
in partnership with the Haitian mobile company, Digicel.
The main goal of this evaluation was to assess project impact on beneficiaries (target families
as well as participating shops) and the effectiveness and relevance of the electronic food voucher as
implemented by CARE. Both qualitative and quantitative methods were applied. Qualitative data
was collected from the main stakeholders of the project (CARE staff, Digicel and Transversal staff,
USAID, community leaders etc.) via a series of focus group discussions. Quantitative data was
collected from beneficiary families, selected shops and non-beneficiary families (control group) for
comparison purposes.
Results from benefitting families were compared with a baseline study. However, data from
shops were collected in a retrospective manner (ex-post and ex-ante data), as no formal baseline
existed. The evaluation was conducted by a team composed of an external consultant supported by
two external monitoring and evaluation experts.
1.1 Beneficiary outcomes
Women remained predominantly head of the household from baseline (69%) to final
evaluation (65%). On average, interviewed household size in the baseline study was 5.6 while it
accounted for 4.9 in the final evaluation.
Increasing food availability was one of the expected outcomes of the project. The proportion
of benefitting families complaining lack of food over the last six months dropped by approximately
51% from the baseline to final evaluation. Moreover, 65.2% of the control group respondents,
suffered lack of food, while only 46.5% of the beneficiary respondents did. The project also intended
5
to support retention of assets of the beneficiary families with findings confirming an increase in
livestock and land retention.
The results revealed a positive shift in the possession of poultry with beneficiaries declaring
possession of between five (5) and 15 poultry, increased by 12 % from baseline. Comparisons with
the control group exhibited a positive difference in favor of the beneficiaries. Findings also showed
that after the project, 11% of the beneficiaries owned more goats and pigs than they did before. This
positive change was confirmed by the proportion of beneficiaries who declared having sold their
animals before and after the project.
Results also showed increased land retention, improvement of personal finance and shift of
debt from basic needs to fund education. After the project, 5% of direct beneficiaries claimed
retention of more agricultural land. In addition, from baseline to end of the project, the proportion of
beneficiaries who were in debt dropped by 33%. Meanwhile debt of control group remained 8%
higher than the treatment group.
Lastly, the final evaluation showed a positive impact on enabling education of children.
Approximately 20% of beneficiaries declared that their children would not be able to go to school
were it not for the project. Moreover, among the beneficiaries, 56.3% mentioned that savings on
food expenditures contributed, among others, to education of their children.
1.2 Program effectiveness
Selection of beneficiary families was drawn from a participatory approach involving
community representatives such mayors, ASEC and CASEC, religious and informal leaders in the
communities. Though absent of1
, criteria adopted by CARE for beneficiary family selection was in
conformity with the World Food Program definition of food insecurity. Bias from the selection
committee was regulated by a systematic verification process but not without impact on program
operations.
According to beneficiaries, the food basket was appropriate with majority indicating a
preference for the form of intervention as provided by CARE. In general, more than 90% were
satisfied with the project. Overall levels of satisfaction with the food basket was very high, though
there were some concerns raised by community leaders regarding the short period of the program.
1
Anthropometry provides portable, universally applicable techniques for assessing the size, proportions and composition of the
human body. It reflects both health and nutritional status and predicts performance, health and survival. Monitoring indicators:
increased percent of eligible children in growth monitoring/promotion and increased percent of children in growth promotion
program gaining weight in past 3 months. Such indicators are important but less relevant in emergency food programs.
6
who voiced that they felt a longer period was required to change nutritional behavior and guarantee a
long-time improvement in the food security of the most vulnerable population.
Although encountering an initial learning curve in the first months, in overall, the electronic
food voucher as implemented by CARE was successful, facilitating confidentiality, flexibility to
food acquisition and monitoring of the process. CARE partner, Digicel, designed and managed the e￾voucher software including: electronic registration of the beneficiaries, voucher issuance,
redemption and merchant payment. This system reduced fraud, logistic burdens and distortion of
local market channels. During the first four months of implementation, average total errors in the
system was 5443 but dropped to 1466 in the last two months. Lack of coordination between partners
and tight project schedules including logistics were the main issues.
Evaluation findings revealed improvement of merchant businesses and overall local
economy. Shop managers revealed improvement of their organizational capacities and reinforcement
of the supply chain with suppliers, including those for local products. All shops but one had a bank
account and managers reported an increase by 13.6 % of tax declaration compared to baseline. In
addition, more than 90 % of retailers in the program declared that their local product stock increased.
Also compared to baseline, an increase of 4.5 % of merchants claimed cultivating local products to
sell as coping strategy. However they complained that local production was less elastic to respond to
demand due to production seasonality. In addition smaller retailers and local traditional markets
activities reported decreasing in activity around voucher redemption days.
1.3 Recommendations
1.3.1 Main recommendations
Sustainability options: In order for emergency food response to improve beneficiary
conditions over a longer period it should be coupled with long term interventions. Agencies
implementing short term food assistance programs need to leverage their programs with other
stakeholders in order to complement respective interventions. This can be accomplished through
cooperation on other programs that aim to sustain local production by providing economic incentives
(grants, technical assistance) or those that link local producers and food retailers.
Reinforcing food insecurity assessment: The monitoring and evaluation system can be
made more illustrative by providing case studies to local community leaders. As well, a systematic
verification of preliminary selected beneficiaries prior the beginning of the program would improve
7
efficiency. Even if not currently in the emergency food program monitoring system, using
anthropometric measures would provide additional insights.
1.3.2 Secondary recommendations
According to beneficiaries and some stakeholders, the duration of project implementation
was too short to have a sustainable impact on vulnerable families. In addition, training sessions
would benefit from being be longer, practical and more illustrative. Implementing agencies should
use mass communication (radio and talk show) to inform on the population on the program. As well,
the implementing agency should consider at least one information and technology staff for
knowledge transfer purposes. Finally, the technical partner of electronic food voucher should
consider scale up strategy and technologies that both cut cost and are easy to be used by vulnerable
population with low literacy levels.
8
2. Introduction
2.1 Background
The devastating earthquake of January 2010 has increased the risks of food insecurity within
poor rural households in the Grande-Anse department of Haiti. The supply of food has decreased due
to deterioration of already poor infrastructures and limited market linkages. The availability of food
on a per person ratio also decreased due to an increase of households members as a consequence
from influx of displaced persons from Port –au-Prince.
In November 2010, Hurricane Thomas worsened the vulnerability of the Grande-Anse
population by destroying assets from agriculture and fishing in the department. In the commune of
Moron, households reporting losing 90 % of agricultural goods. Two months later, food prices for
basic staples had increased up to 21% according to the national commission for food security
(CNSA). Rice and local corneal prices climbed sharply respectively by 21 % and 20% while cooking
oil prices increased by 9%. Those two events combined forced populations living with less than $ 1
a day to allocate more than 60 % of its income to food.
Soon after, a cholera epidemic broke out in Haiti, with the heaviest number of cases reported
in the Grande-Anse. Studies revealed that the hurricane and cholera outbreak, along with rising food
prices, greatly increased family vulnerability. In addition, an increased number of orphans
augmented the risks of health issues for children including malnutrition.
To tackle the aforementioned problem, CARE submitted a proposal to USAID/Food for
Peace addressing food insecurity concerns for populations in nine communes of the Grande-Anse
department. Approved by USAID, this project named “Electronic Food Voucher program – Kore
L’avni Nou”, aimed to support 12,000 vulnerable families (72,000 individuals) with a $50 (2,000
HTG) monthly voucher over a six month period. In addition, 68 local shops were approved to supply
beneficiaries with the following key commodities: Rice, vegetable oil, maize and beans, forming a
minimum food basket and required kilocalories per family per month.
Before implementing the project, a baseline study was conducted in January 2012 to collect
information on food availability, current household situations, retention level of net assets, negative
coping strategies, etc. Upon termination of the project2, CARE Haiti commissioned an external
evaluation of the core aspects of the “Kore L’Avni nou” project. The goals of this evaluation were:
2
 As per approval from USAID, the final Evaluation of Kore L’avni Nou does not take into consideration phase 2 (cost-extension) of
the project in response to hurricane Sandy aftermath in Grande-Anse.
9
 Identify and evaluate program outcomes, intended and unintended, positive or negative
and impact on beneficiaries versus non-beneficiaries
 Overview data collected on a regular basis and present performance monitoring (strengths,
weaknesses), triangulate with final evaluation collected data for better understanding
 Determine the adequacy of program design to meet the identified needs in the target area
 Compare final conditions with baseline findings and present persisting gaps in coping
strategies used by the households to mitigate vulnerability
 Relate lessons learned to appropriate stakeholders
 Identify key obstacles preventing a correct program execution
 Identify key successes and recommendations for future consideration
2.2 Methodology
The external evaluation of the CARE Food Voucher program was conducted by using both
qualitative and quantitative methodologies. The methodology and instruments to conduct the study
took into consideration the baseline study findings and inputs from CARE staff. The complete
evaluation was carried out by an external team composed of a consultant and two monitoring and
evaluation experts. A team formed of one experienced supervisor, nine experienced enumerators and
nine local guides from the nine different communes conducted data collection. Field data were
collected on a mobile device (tablets) and automatically uploaded on an online application platform
developed by Diagnostic & Development Group. Then data were exported to SPSS for statistical
analysis.
2.2.1 Quantitative methodology
A comprehensive survey addressed Food Voucher program “Kore L’avni Nou” beneficiary
families, non-beneficiary families and beneficiary merchants/shops. Units of analysis were selected
through a probabilistic sampling approach. The overall beneficiaries (12,000 families and 683
 local
shops) constituted a sampling frame for drawing out a sample of family beneficiaries and local
shops. A proportional sampling size method based on the principle of spatial and population
3
 As per program management and reported elsewhere in the reports, 2 merchants/shops were disqualified for fraudulent activities and
therefore have not been taken into consideration during sampling.
10
representativeness with a Quota method) used to choose the samples. Hence, the number of
beneficiaries and shops that was selected in each zone was determined by the quota of beneficiaries
and shops it holds.
The sample size of beneficiary groups was calculated in order to ensure a 95%-confidence
level and a 5%-margin of error. For a 12,000- beneficiary population, the sample size to guarantee
the aforementioned requirements should be of 374 (considering a finite population). This number
was then rounded up to 450. Due to budget and time constraints, the control group (non beneficiary
families) was arbitrary chosen to be 60% of the sample beneficiaries or, 270. Finally, a third of the
total 68 local shops constituted the sample for merchant shops approved to supply the food basket to
beneficiaries.
To select the control group sample, the following strategy was adopted. Once in his/her
assigned commune(s), the surveyor began at the first street corner and entered the first “Kore L’Avni
nou” project non-beneficiary household to start the survey. Each successive fifth household was then
surveyed. If a household was not eligible, the next was chosen, and so on. This strategy was repeated
until the projected sample size was attained.
2.2.2 Qualitative methodology
Complementary to quantitative techniques, focus groups and semi structured interviews were
conducted. Focus groups were held with community leaders at the commune level. Interviews were
conducted by key informants including staff and partners of the project.
2.2.3 Sample size and data collection
Field data collection took place in nine (9) targeted communes of the Grande-Anse
department from January 18th to January 26th for the three (3) target groups : benefitting families,
participating shops and non-benefitting families with questionnaires composed of both closed and
open questions. The following issues were faced during data collection:
 Some had migrated out of the locality.
 Some passed away but the food was allocated to a relative or another family member.
 Some beneficiaries were difficult to locate as are known better by their nicknames
As can be seen from Table 2.1, in all but one commune, Les Irois, more than two shops
participated in the evaluation survey, therefore target samples size were attained. This was due to
11
the sample list being increased for each commune, in the case that some shops could not be reached.
In the case of Les Irois, the surveyor did more than actually needed.
Table 2.1 Shops sampling
Commune Number %
Sample size
target
#
reached
Roseaux 8 12,3 2 2
Dame Marie 7 10,8 2 2
Pestel 15 23,1 5 5
Chambellan 4 6,2 1 1
Moron 3 4,6 1 1
Beaumont 9 13,8 3 3
Anse d'Hainault 5 7,7 2 2
Corail 8 12,3 2 2
Les Irois 6 9,2 2 4
Total 65 100 20 22
For the beneficiary sample, in all communes but Beaumont where the number of respondents
reached were less than planed (3), samples were as projected in the revised technical proposal (Table
2.2).
Table 2.2 Beneficiairies sampling
Commune Number % Sample
#
reached
1 Anse DAinault 1 492 12,4 56 56
2 Beaumont 983 8,2 37 34
3 Chambellan 1 138 9,5 43 43
4 Corail 1 383 11,5 52 55
5 Dame Marie 1 721 14,3 65 65
6 Les Irois 1 058 8,8 40 42
7 Moron 1 075 9,0 40 40
8 Pestel 1 866 15,6 70 72
12
9 Roseaux 1 284 10,7 48 48
Total 12 000 100 451 455
As relates to the non-beneficiary group (control group), except for Beaumont where three
respondents less than projected were interviewed, the sample group size remained as projected
(Table 2.3).
Table 2.3 Non Beneficiaries sampling
# Commune Number % Sample # reached
1 Anse DAinault 1 492 12,4 33 33
2 Beaumont 983 8,2 22 19
3 Chambellan 1 138 9,5 26 26
4 Corail 1 383 11,5 31 34
5 Dame Marie 1 721 14,3 39 39
6 Les Irois 1 058 8,8 24 27
7 Moron 1 075 9,0 24 24
8 Pestel 1 866 15,6 42 46
9 Roseaux 1 284 10,7 29 35
Total 12 000 100 270 283
Moreover, “no-answer” analysis showed that, in general, for all the questions there were no
more than three (3) missing answers, hence we may conclude that for the overall evaluation study,
the results were as expected.
2.3 Limits of the evaluation
Primary limitations of this evaluation study are linked to non existence of a control group in
the baseline study, which does not allow a “difference in difference” approach to measure the net
impact of the project. In addition, it did not take into account anthropometrics indicators such weigh,
height, body mass index, etc,. which are very important in food security program, however, nutrition
improvement aspect was not part of this emergency intervention.
13
Furthermore, the selected local merchants/shops were not included in the baseline study. This
challenge was solved by using a retrospective approach however, with such an approach; the data
may have been affected by a so called memory bias problem.
Overall, data collection process was challenged by issues such as:
 Some had migrated out of the locality.
 Some passed away but the food was allocated to a relative or another family member.
 Some beneficiaries were difficult to locate as are known better by their nicknames
 Some stakeholders refused to participate in focus group.
2.4 Organization of the report
Findings of the CARE Food Voucher program evaluation are herewith reported in three
sections. The first section presents socio-demographic profiles of families, and discussions of how
the program impacted their livelihood and coping strategies. It also assesses the impact of the
program on merchants/shops and local economy. The following sections focus on an overview of
collected data and the monitoring and evaluation system performance, the effectiveness and
relevance of the program, key lessons learned, the challenges faced by the project, and finally key
recommendations.
3. Evaluation results
As per the terms of reference, this section focuses on the impact of the project on
beneficiaries versus non-beneficiaries, and comparison of final evaluation findings with those of the
baseline.
 3.1 Evaluation of impact and comparison with the baseline study
As aforementioned, the impact evaluation is twofold: impact on beneficiaries, hereafter
referred to as treatment group, and impact on local merchants/shops. This subsection of the report
presents impact analysis on beneficiaries by comparing the final evaluation results for the
beneficiary group with those of the baseline study. It also assesses the impact of the “Kore L’Avni
Nou” project on selected local shops in a retrospective fashion (before vs. during the project course).
14
3.1.1 Socio-Demographic profile of the respondents
The profiling of the households entails household size, sex of household head, age, and
external support received. The mean household size was 5.6 in the baseline study while it accounted
for 4.9 in the final evaluation. The mean of household size for control group (non beneficiary
families), was 4.9 which is closer to the final evaluation. The results also showed that from the
baseline to final evaluation, on average, the beneficiary household head was 63 and 42.5 years old
respectively, while for the control group, the mean age of household head was about 34 years. (See
table 3.1).
Table 3.1 Mean household size and age of household head
Treatment group Control group
Baseline Final evaluation
Household size 5,6 4,9 4,9
Age of household
head
63 42.51 33.76
From the baseline to final evaluation, a significant number of women remained as head of the
household. The baseline and final evaluation found that female headed households respectively
represent 69% and 65 % respectively. The control group had a percentage of men head of
households of 37.3 %. (See table 3.2)
Table 3.2 Distribution of household head by sex
Treatment group Control group
Gender
Baseline Final evaluation
Male 31,5% 35,1% 37,3%
Female 68,5% 64,9% 62,7%
Total 100% 100% 100%
For statistical comparison between the treatment and the control groups, nine (9) non￾beneficiaries reporting having received external support from other project, were excluded from the
analysis. Furthermore, the validity of the analysis was considered by minimization of the influence
15
of other factors such supports from family, friends, etc. Assessment of those factors revealed
existence of no significant impact difference of external assistance on the two groups, which allow
us to hypothesize that they are largely comparable by only the outcomes of the Food Voucher
program (Table 3.3). In fact, the percentage of the “Kore L’avni Nou” beneficiaries receiving such
support is about 0.6 point less than those of the control group.
Table 3.3 External support
Answer Treatment group Control group Difference
Yes 8,80% 9,40% -0,60%
No 91,20% 90,60% --
Total 100% 100% --
3.1.2 Impact on food security for beneficiaries
The program provided significant improvement of food access to targeted vulnerable
families. The impact access to food improved by 51 %. Simultaneously, household complaint of lack
of food dropped from 97.7 % , before the program, to 46.5 % , after implementation. Increasing
access to food was one of the expected impacts of the project on the beneficiaries. It is worth
noticing that food insecurity has4
 been a chronic issue in this region. Before the project, respondents
reported insecurity related to food availability over the past 12 months. This improvement of food
access was reported just after the six months of project implementation.
In absence of the CARE Food Voucher program, food insecurity in the communes would
remain significant. The non-beneficiaries of the program reported a lack of access to food of about
65, 2 % at end of the program while it was only 46.5 % for beneficiaries. This positive impact on
beneficiaries is confirmed by this 17.7% difference between the control group respondents and those
from the treatment group. (See table 3.4)
4
www.fews.net/FoodInsecurityScale
16
Table 3.4 Impact on food availability
Treatment group
Lack of
food
Baseline Final evaluation Difference Control group Dif. (Fin. Eval. –
Cont. group)
No 2,3% 53,5% 51,2% 35,8% 17,7%
Yes 97,7% 46,5% 65,2% --
Total 100,0% 100,0% 100,0% --
3.1.3 Effect on beneficiaries coping strategies
Analysis of food insecurity was also assessed through the mitigation strategies of households.
Respondents were asked about strategies used to maintain their livelihoods over the 30 days
preceding the survey. Questions related to strategies such decrease of food ration volume and
quality, food credit and search of external assistance for food. Comparative analysis with baseline
addressed whether the treatment group faced this situation.
Evaluation of the project on food availability also considers the distribution of the
respondents and whether they never faced the aforementioned situation (figure 3.1). Comparison
with the baseline results revealed that after the project, 12.7% of beneficiary families relied on help
from family or neighbors to feed the household, 6.5% went without eating for a full day, 5.9% ate
unwanted food or food from an immature/early harvest, and 1.8% reduced the number of daily meals
less than they did before the project execution (Table 3.5)
17
Figure 3.1. Coping strategy as percentage of population (control group, baseline and beneficiaries) who never
faced this situation
Table 3.5 Impact on coping strategies
Treatment group Control
group
Dif. (Fin. Eval. – Cont.
group)
Coping strategy
Never faced this
situation
Baseline Final evaluation Difference
Eat unwanted food or
food from an
immature/early harvest
9% 14,9% 5,9% 9,9% 5,0%
Borrow food 45% 44,6% -0,4% 40,5% 4,1%
Buy food on credit 7% 7,9% 0,9% 9,9% -2,0%
Rely on help from
family or neighbors to
feed the household
22% 34,7% 12,7% 32,1% 2,6%
Consume seed stock 26% 24,6% -1,4% 17,5% 7,1%
18
Table 3.5 Impact on coping strategies
Treatment group Control
group
Dif. (Fin. Eval. – Cont.
group)
Coping strategy
Never faced this
situation
Baseline Final evaluation Difference
Leave home to beg 75% 73,4% -1,6% 65,0% 8,4%
Reduce the size of
food rations
5% 5,5% 0,5% 3,3% 2,2%
Reduce the number of
daily meals
3% 4,8% 1,8% 2,2% 2,6%
Restrict consumption
from some members in
order for others
(children) to have
enough to eat
33% 19,8% -13,2% 15,0% 4,8%
Go without eating for
a full day
6% 12,5% 6,5% 8,4% 4,1%
Work for food 47% 36,9% -10,1% 30,3% 6,6%
Give family salty
vegetables, salty tea,
or salty coffee as only
alternative
20% 15,2% -4,8% 12,0% 3,2%
However, deterioration may be observed for the following:
 Restrict consumption from some members in order for others (children) to have enough to eat
(-13.2%)
 Work for food (-10.1%)
 Give your family salty vegetables, salty tea, or salty coffee as only alternative (-4.8%)
 Leave home to beg (-1.6%)
The results also showed that in all but one case fewer beneficiaries faced such situations than
did their peer of the control group (with positive differences varying from 2.2% to 8.4%).
19
3.1.4 Benefits on households assets
The electronic Food Voucher program was expected to improve retention of net assets by
beneficiaries. As with many rural areas in developing countries, households assets are mainly
composed of livestock and agricultural land.
3.1.4.1 Impact on livestock asset
Project impact on the beneficiary livestock was gauged by comparing the ownership of
poultry, goat/pigs and cows before and after the project.
Table 3.6 shows a remarkable shift in the number of poultry owned by beneficiaries after project
execution, though it is not obvious to claim that more beneficiaries had access to livestock after the
course of the project (about 1% difference between those with zero (0) poultry before and after the
project).
About 12% of the family beneficiary respondents declared to own between one (1) and five
(5) poultry less than they did before the project, while approximately the same proportion affirmed
to own between five (5) and fifteen (15) poultry after the project, more than they did before.
Comparisons with the control group exhibit positive difference in favor of the beneficiaries.
Table 3.6 Impact on ownership of poultry
Treatment group Control group Dif. (Fin. Eval.
– Cont. group)
# of poultry Baseline Final evaluation Difference
0 58% 57% -1% 61% -5%
1--2 17% 8% -9% 10% -1%
3--4 14% 11% -3% 10% 2%
5--10 9% 20% 11% 18% 2%
11--15 1% 2% 1% 1% 2%
15 and more 1% 1% 0% 1% 1%
Total 100% 100% 100%
20
Concerning ownership of goat/pig, it is evident to assert that, based on the data, more
beneficiaries owned goat/pork after the project execution than before. From table 3.7 we learn that it
appears that about 11% of beneficiaries who did not possess goats/pigs before the project start, with
in an inverse situation after project implementation. Furthermore, one can observe that, in general,
after the project, the proportion of beneficiaries who owned these animals was higher than before its
start.
Table 3.7 Impact on ownership of goat/pig
Beneficiaries Control group Dif. (Fin. Eval. –
Cont. group)
# of
goat/pig
Baseline Final evaluation Difference
0 81% 70% -11% 76% -6%
1 7% 8% 1% 4% 4%
2 5% 11% 6% 8% 3%
3 4% 4% 0% 3% 1%
4 1% 3% 2% 4% -1%
5 and more 2% 4% 2% 5% -1%
Total 100% 100% 100%
The same dynamic may be observed for the ownership of cows (see table 3.8).
Table 3.8 Impact on ownership of cows
Treatment group Control group Dif. (Fin. Eval.
– Cont. group)
# of cows Baseline Final evaluation Difference
0 93% 89% -4% 91% -1%
1 6% 7% 1% 5% 1%
2 1% 3% 2% 2% 1%
3 and more 0% 1% 1% 2%
Total 100% 100% 100%
21
The positive dynamic observed on the ownership of animals is reinforced by considering the
proportion of beneficiaries who declared having sold animals over the last three months prior to data
collection. The results of the final evaluation showed that 3.6% and 3.2% of beneficiaries reported
having sold less transport animals and cows than they did for the baseline study (Table 3.9).
Regarding the selling of goat/pigs, the situation after the project is not greatly different than before
implementation. Comparisons with the control group respondents reveal that these latter were more
prone to sell their animals than were the beneficiaries.
Table 3.9 Impact on selling animals
Treatment group Control group Dif. (Fin. Eval. –
Cont. group)
Selling animals Baseline Final evaluation Difference
Transport Animal 4% 0,4% -3,6% 0,7% -0,3%
Beef 5% 1,8% -3,2% 2,6% -0,8%
Goat/Pork 7% 7,5% 0,5% 8,9% -1,4%
3.1.4.2 Impact on beneficiary land assets
Impact on beneficiary net assets was also gauged considering whether these latter have sold
agricultural land during the life of the project. However, the results of the baseline study showed
that was not a concern, since 90% of the beneficiaries then declared not having sold such assets over
the last three months prior to the project start (Table 3.10).
 The results of the final evaluation show that 5% of the beneficiaries reported not having sold
agricultural area over the last three months more than they did in the baseline. It is not obvious that
this positive change might be attributed to the project, since the proportion of the control group
respondents that gave such an answer was essentially the same as for beneficiaries. However, it is
worth noting that land ownership in general is an issue, therefore it would be normal to not see a big
difference between the two groups.
Table 3.10 Impact on selling agricultural land
Treatment group Control group Dif. (Fin. Eval. –
22
Cont. group)
Selling agricultural
area
Baseline Final evaluation Difference
No 90% 95% 5% 94% 1%
Yes 10% 5% -- 6%
Total 100% 100% -- 100%
3.1.5 Impact on beneficiary’s personal finances (debt and borrowing)
Respondents of both studies (baseline and final evaluation) were asked whether they owed money
during data collection periods. Comparison of both results showed a significant decrease (33%) in
the proportion of beneficiaries who owed money before the project execution as compared to those
who did so after the project. The results also revealed that 8% of control group respondents owed
more money than did those of the treatment group. On average, the beneficiaries declared owing
3,821.189 gourdes against 5,608.72 gourdes for the non-beneficiaries5
. (See table 3.11)
Table 3.11 Impact on debt
Treatment group Control group Dif. (Fin. Eval. –
Cont. group)
Debt Baseline Final evaluation Difference
Yes 86% 53% -33% 61% -8%
No 14% 47% -- 39% --
Total 100% 100% -- 100% --
Considering the main reasons of borrowing money, as could be expected, the proportion of
beneficiaries who affirmed having borrowed money to buy food after the project was lower (8%
difference) than those who did so before its implementation. The results showed that money was
borrowed for more long term investments such education, housing and re-investments (for example
seeds for agriculture) (See table 3.12). This shift of allocation of resources is logical to the fact that
the program responded to primary needs (food) of beneficiaries, therefore it allows them to focus on
higher ranked needs.
5
42 gourdes = 1 $USD
23
Table 3.12 Reasons for borrowing money
Reasons for
contracting debt
Baseline Final evaluation Difference
Food 45% 37% -8%
Health 23% 21% -2%
Education 16% 23% 7%
Housing 0% 3% 3%
Re-investment 2% 11% 9%
Marriage, funeral 9% 3% -6%
Other 5% 1% -4%
Total 100% 100%
3.1.6 Impact on beneficiary children education
To measure the impact of the project on education, the respondents (treatment and control
groups) were asked whether they had schooled children during the project period. Approximately
27% of the beneficiaries answered yes, while about 15% of non-beneficiaries gave the same answer.
Thus, nearly 12% of the treatment group respondents sent children to school during the project
execution more than did those of the other group. (See table 3.13).
Table 3.13 Children schooling during project the project course
Answer Beneficiaries Control group Difference
Yes 27,1% 15,3% 11,8%
No 72,9% 84,7%
Total 100% 100%
To confirm that this difference was attributable to the project, beneficiaries were also asked
whether they had children who would not have gone to school were it not for the project. The
proportion of positive answers obtained was about 20.1%. Furthermore, among beneficiaries, 56.3%
mentioned that savings on food expenditures contributed, among others, to education of their
children.
24
3.1.7 Access to mobile phone
 From the baseline study to final evaluation, the gap of access to phone communication
decreased. This unintended change is very important considering the role in access to
communication for rural and vulnerable households. Use of the electronic food voucher exposed the
target group to phone technology; in fact 75 % of respondents confirmed access to phone. Table 3.14
confirms this assertion by showing a positive change of 29%. However, this result is mitigated when
compared to the control group with an observed negative difference of 10% (treatment group –
control group). Thus, this positive change might be attributed to a more general “dynamic” of
communication market expansion.
Table 3.14. Access to mobile phone target groups
Treatment group Control
group
Dif. (Fin. Eval.
– Cont. group)
Acces to phone Baseline Final evaluation Difference
Yes 46% 75% 29% 85% -10%
No 54% 25% -- 15% --
Total 100% 100% -- 100% --
3.1.8 Effect on shops and on local economy
In addition to improvement of living conditions of the most vulnerable populations,
distributing vouchers instead of physical commodities, also boosted the local economy. One of the
most important features of “Kore L’Avni Nou” is the use of local merchants/shops, mostly retailers,
as distributors/procurers. That strengthened shop capacities, injected money into the local economy,
created jobs, increased of the food demand and, to a lesser extent, reinforced local products though
most of the basket goods are imported commodities. Hence, this part analyzes impacts of the project
on the “beneficiary” merchants/shops6
 and local economy.
The survey on shop owners (merchants) was conducted to capture data on the impact of the
project on local economy, shop managerial capacities and performance. These data were pooled with
6
 The project did not have specifically target activities aimed at merchant/shop capacity improvement other than trainings on the use of
technology and introducing record-keeping (receipts).
25
administrative data to perform the analysis. A breakdown of the survey conducted in tabular and
graphical forms is presented in the following section.
3.1.9 Impact on shops' managerial capacity
Information regarding the change in the shops ‘internal management’ during the project is
summarized in figure 3.2.
From the 22 interviewed shops, only 50% had an account book to register their transactions
before the project. This situation did not seem to be improved during the project course. In spite of
this lack of accounting system, all were compelled to use receipts to record every purchase/sale
transaction during the project, while only 77.3% had used it before the project. Concerning bank
account, all shops but one had a bank account while just 50% had before the project implementation.
It also appears that, according to administrative data, all shops were tax compliant during the project
against 63.6% respectively before.
Figure 3.2. Impact on shop managerial capacities
3.1.10 Impact on amount of money in local economy
Table 3.15 reports the amount of money injected directly into the local economy to boost
spending, especially of the most vulnerable people. In fact, it may be observed that more than 141
million Gourdes have been injected into the local economy through the vouchers program. That
represents, on average, more than 15 million Gourdes per commune or more than 2 million Gourdes
per shop. Considering that shops are mostly retailers, the amount injected is relatively important.
26
However, this injection would have a real effect (at mid and long term) on local economy only if it
affected local product demand and, above all, the production sector.
C o m m u n e N u m b e r o f r e d e m p t io n s A m o u n t
A n s e d \ H a in a u lt 8 7 9 6 1 7 5 8 2 6 0 0
B e a u m o n t 6 8 9 5 1 3 7 8 6 2 0 0
C h a m b e lla n 7 4 1 6 1 4 8 2 2 9 6 7
C o r a il 6 9 7 5 1 3 9 3 6 6 2 0
D a m e M a r ie 9 9 2 5 1 9 8 4 8 0 0 0
L e s I r o is 6 2 5 1 1 2 5 0 2 0 0 0
M o r o n 5 7 2 4 1 1 4 4 4 2 0 0
P e s t e l 1 0 9 5 8 2 1 9 1 0 6 0 0
r o s e a u x 7 6 5 8 1 5 3 0 8 8 0 0
T o t a l 7 0 5 9 8 1 4 1 1 4 1 9 8 7
 T a b le 3 . 1 5 . A m o u n t o f m o n e y I n je c t e d in t o lo c a l e c o n o m y *
*Administrative data: extracted from the system7
3.1.11 Effect on local production and employment
In response to money injection, it seems that the demand of local products increased during
the project execution. Indeed, based on data analyzed on table 3.16, among the shops, 95.5%
(63.6%+31.8%) agreed or completely agreed that their local product stock increased during the
project. The most products sold were: rice, maize, beans and food ingredients such as oil, spice, etc.
Corn and local beans were the two most demanded local products during the project. However,
according to the shop owners, seasonal harvesting of local products was an important problem,
which complicated the adjustment of the supply to the demand.
In addition, all the respondents claimed increased sales of local products during the project.
Moreover, 77.3% and 13.6% completely agreed and agreed respectively that their profits increased
during the project, while only 4.5% (one) completely disagreed.
Completely agree Agree Somewhat agree
Local Product Stock increased 63,6 31,8 0
Sales increased 77,3 22,7 0
Profits increased 77,3 13,6 4,5
Table 3.16. Impact on local market (%):
7
 Extract from the « Liste de vendeurs_68 points de vente/Stores 24 dec_2012 »,Project document from CARE.
27
Although the project did not focus on job creation, all the shops affirmed having created jobs
or increased the number of employees to respond to the increasing demand and activities during the
project. This is in contrast to the fact that only 68.2% of shop owners (Table 3.17) claimed they had
personnel at the beginning of the project. However, some of them were skeptical they might be able
to keep these jobs after the project. The project implementing partner, Digicel, claimed to hire
additional workers over the course of the project.
Before the project After the project Difference
Job Creation 68,2 100 31,8
Tabble 3.17. Effect on Job creation (%):
3.1.12 Effect on shop coping strategies
Merchants have developed coping strategies to invest or grow their business against a
downturn of circumstances. Coping strategies developed by shop owners were assessed with respect
to how often they were adjusted by shop owners. In order to assess this fact, data was collected
before and after the project. Figure 3.3 and Table 3.18 provides a complete breakdown of data
collected related to each coping strategy.
Figure 3.3. Effect of the project on shop coping strategy (%)
28
Before the project After the project Difference
Borrowing Money 27,3 18,2 9,1
Buy Merchandise on Credit 36,4 31,8 4,6
Cultivate some products to sale 18,2 22,7 -4,5
Credit from particular with high rate 18,2 13,6 4,6
Selling animals 27,3 13,6 13,7
Selling agricole Area 0 9,1 -9,1
Table 3.18. Effect on Coping Strategies (%):
After the project, among the six (6) identified coping strategies, four (4) moved in the
expected direction: borrowing money, buy merchandise on credit, credit from particular and selling
animals. Based on these findings, the project has reduced “selling animals” by half, from 27.3% to
13.6%. As regards credit, “borrowing money” was lowered by 9.1 %, “Buy merchandise on credit”
and “Credit from particular”, by 4.6% each. Additionally, during the project one shop owner (4.5%)
sold products from his harvest. In this case the shop owner is at the same a local producer and
distributor through his shop. Finally, two sold some agricultural lands to invest and grow their shops.
3.2 Overview of data collected on regular basis and presentation of
performance monitoring
This subsection analyses the monitoring and evaluation system (M&E) of the Food Voucher
program, and presents the different type collected by the associated staff. It also maps the flow of
information generated by this system and how this information gets to key stakeholders.
3.2.1 Monitoring and evaluation system
The assessment revealed that M&E of the project tackled four main tasks: targeting of
beneficiary eligibility, ensuring the quality of food delivery, service performed by merchants/shops
and follow up on use of assistance by beneficiary (post distribution surveys). This section of the final
evaluation assesses how the on-going monitoring system provided continuous feedback on the
project evolution, the main challenges faced by M&E staff during the course of the project, etc.
Some improvement points will also be presented. The analysis that will be presented based on the
results of the M&E staff focus group.
Following registration in the system and the first voucher redemption, the Monitoring and
Evaluation unit undertook a systematic revision of eligibility criteria. Each month, 20 % of the
29
overall 12,000 was selected randomly in the commune to assess eligibility criteria from February to
July 2012. After this period, the verification ratio decreased by 10% due to lack of human resources
and need to correct duplication cases or voucher swapping problems by tracing individual
beneficiaries and addressing the problem.
According to administrative data, 10 families who did not comply with vulnerability criteria
were originally selected and registered. They were revealed and removed from the project list being
replaced by other families. The monthly verification process was followed by a correction process
involving activation and deactivation of vouchers by transversal technicians.
30
3.2.2 Data collected by the monitoring and evaluation staff
CARE electronic Food Voucher management shortcutted the channel of command to
improve efficiency to implement decision making process. Information produced by the M & E
Execution of information decision
Flow of production of information demanded
Fig 3.4 Monitoring and Evaluation information Flow
31
Manager is treated by the Program Coordinator and actions resulting from the analysis are
implemented by the Project Manager who is in direct contact with field agents (Figure 3.4).
Monitoring and Evaluation produces a monthly report including four surveys: commodity market
price, post distribution survey analysis, eligibility criteria verification surveys, and food quality
random verification and analysis (Table 3.19). These data are collected by field agents under the
direction of the M & E field officers. These officers reported that lack of resources (Field Officers
are not totally dedicated to M & E as they also support beneficiary duplication and voucher
swapping clearance) and administrative burden (conflicting agenda of field agents with
administrative task, access to fuel, logistics were obstacles in data collection as well as overall field
operations).
Analyzing information system conflicts of interest rose in data collection and decision
implementation related to community mobilizer roles (Principal agents’ issue). Those mobilizers are
in charge of beneficiary supervision as well, they execute decisions i the field. Therefore, there may
be biased in reporting field feedback to top project manager or field office managers. The CARE M
& E manager reported that counter surveys and focus groups were conducted to mitigate this
potential issue. However we cannot exclude the probability.
Table 3.19 Reports produced with data collected by M&E staff
Reports
Commodity
market price
report
Post distribution
Report
Eligibility criteria
matching report
Quality product
report
Description
& objectives
Assesses the
basket commodity
prices on the
market provided
by the shops, and
helps to maintain a
fair price by the
shops. Retailers
were asked
mandatorily to
This report aims to
document the use of
beneficiaries of the
food.
Provides
information on the
profile of
preliminary selected
beneficiaries by
community leaders.
It helps to ensure the
quality and respect
of the vulnerability
criteria of the
Assesses
inventory stock
of the shops to
ensure the
quality and
availability of
products for
targeted
beneficiaries.
32
Reports
Commodity
market price
report
Post distribution
Report
Eligibility criteria
matching report
Quality product
report
publicize their
price (we can
analyze those
reports)
targeted group.
Methodology
Random selection of
20 % of the
preliminary
beneficiaries
selected by
community leaders
from each
commune. This
verification process
of eligibility criteria
held from February
2012 to July 2012,
then the sample
decreased to 10 %.
3.3 Relevance and effectiveness of the electronic Food Voucher program in
Grand’Anse
This section presents and analyzes the effectiveness and relevance of the electronic Food
Voucher program implemented by CARE in the Grand'Anse region of Haiti. First, this new CARE
service delivery process is mapped out and presented in the context of food distribution program in
Haiti. Assessment of personal satisfaction of beneficiaries and other stakeholders regarding the
project is also presented.
33
3.3.1 Food insecurity in Haiti and Food distribution challenges
The staggering poverty rate in Haiti is associated with food insecurity issues that are severe
boundaries to human capital development. According to World Bank, the gross domestic product
(GDP) of Haiti is ten times lower than average of Latin America. In 2012, the Haitian economy
contracted by 2 % due to low productivity of agriculture and bad weather. The Global Hunger Index
ranked Haiti 77 out 79 for the year 2012 and indicated that 38 % of the population is food insecure 8
.
In aftermath of the earthquake while food insecurity increased, distribution of rations to
vulnerable populations remained challenged. Crowded lines, space for distribution, security for
convoy, safety of distributors and potential beneficiaries were among challenges of several
international agencies9
. Logistic and security issues also increased as relates distance in miles to
reach beneficiaries from Port-au-Prince to other cities.
3.3.2 Description of CARE electronic Food Voucher program
The electronic Food Voucher program was a new form of food distribution to provide
assistance to vulnerable population in nine targeted communes in Grand 'Anse. An electronic
voucher associated with a pin card was issued to targeted registered beneficiaries. This voucher was
credited with an amount of $ 50 valid for a period of 30 days and refilled each month over a six (6)
month period. Beneficiaries then went to a local shop registered in the program to redeem his
voucher. The shop manager used a USSD phone on his mobile phone (*567* voucher number*pin#)
to communicate with Digicel electronic platform “Merchant Pro”. Then an SMS is sent back to
confirm existence of credit for the given beneficiary. This successful message allowed merchants to
deliver a preselected food basket to beneficiaries. CARE personnel then proceeded with field
beneficiary verification leading to reconciliation of food voucher transactions. Later, shops went to a
local microfinance, Fonkoze, with Digicel to access the value of the goods sold. Once verified, the of
beneficiary accounts were credited to restart the food voucher cycle (figure3.5).
8 WFD 2012
9 New York Times, problems with food distribution in Haiti, January 27th 2010
34
Figure 3.5 Program operations
3.3.3 Effectiveness of the program operations
Globally, CARE electronic Food Voucher program operations improved food security of
beneficiaries as compared to traditional distribution. The program operation channels alleviated the
logistical burdens by using local shop supply channels to deliver food to targeted populations. From
beginning to end, the program operations are be reported in 11 steps (Figure 3.5).
3.3.4 Effectiveness and relevance of beneficiary’s selection process
Selection of the Grande Anse department was justified for project implementation after
Hurricane Thomas and based on deteriorating socio-economic conditions of its population also
exacerbated by the influx of internally displaced families as a result of the 2010 earthquake in Haiti.
Grande Anse is an isolated area where the annual revenue (HTG 5,269.00) of independent workers is
35
less than the average of rural areas in Haiti; moreover 82.3 % of communes in Grande Anse have
very weak and weak access to basic services10 . In addition, transportation is very difficult in this
area. The implementation of the project in this region was reasonably justified to recapitalize
farmers, rebuild or strengthen market linkages and assist poor households.
CARE used a participative methodology coupled with defined criteria to select the
beneficiaries. A community committee formed with mayors, community representatives (ASEC or
CASEC), the Citizen Protection office, pastors, judges, teachers and notables of the Communities
helped identify potential beneficiaries. This committee also legitimated the process in the
community but it is reported that in some cases, influent members were biased in their choice either
for political or personal reasons.
Therefore, the second step, data cleaning, was conducted to ensure eligibility of beneficiaries
with respect to relevant criteria. CARE elaborated two sets of criteria: one for vulnerability and
another for extreme vulnerability.
The vulnerable group criteria consisted of:
 People living with less than $2/day,
 People owning less than two livestock and,
 People owning less than 8/100 Cx of land.
In addition to food insecure, the extreme vulnerable criteria were:
 Pregnant,
 Handicapped (Blind, deaf, mute),
 Widowed, orphan,
 HIV or sick,
 Displaced from the 2010 earthquake or Hurricane Thomas.
Aside from the absence of anthropometrics indicators, the above criteria are pertinent and
relevant to define the targeted population based on World Food Program (WFP) definition of food
insecurity. In fact, WFP defines food insecurity as a situation where a household or country is not
able to provide (future) physical and economic access to sufficient, safe, and nutritious food
necessary for living an active and healthy lifestyle.
10 ECVH 437-438 ;Poverty map, Haiti 2004
36
This definition is similar to three components of food insecurity: Availability, Access, and
Use described by World Health Organzation. In addition, WFP considered specific groups such
children, specifically those under five, pregnant women, handicapped as more vulnerable due to their
limited capacity to access food or because they are more likely to be physically or economically
affected by malnutrition. However, to access a good use of food received and mostly for some
people in extreme vulnerability, some anthropometric indicators such body mass measures would be
so useful. Even though, it is necessary to concede that it would occur some additional cost.
Contrary to beneficiary families, the merchants/shops were selected through a competitive
process. CARE conducted an assessment of business capacity, storage, financial capacity and legal
registration of shops. This selection process evicted smaller shops or merchants who did not satisfy
requirements. In addition, the financial gross sales of the shops were subjective even it was coupled
with CARE staff observation of retailer's stock. In order to respond to criteria some shops registered
themselves under the umbrella of shops having legal papers to facilitate product distribution. It can
be questioned whether this cooperation as an unintended effect of the project.
3.3.5 Relevance and Effectiveness of registration mechanism
Registration of beneficiaries was relevant to confirm and monitor vulnerability eligibility
criteria. Registration of beneficiaries allowed CARE to credit personal account of each beneficiary
through a Digicel electronic food voucher program. Therefore, CARE was able to follow up in the
field to access food rations by each registered beneficiary. This process also helped to triangulate
information regarding food availability and delivery by shops, beneficiary voucher redemption and
field data collected by CARE. The system enabled tracing of the information per individual as
registered in different communes.
The registration process of beneficiaries in the electronic system was challenging in the
beginning but significantly improvement during project implementation. The registration process
consisted of matching a voucher number and a PIN number to an individual beneficiary to be
registered in the merchant pro platform. This field operation was implemented by Digicel agents
assisted by CARE community mobilizers. The registration process was far from being efficient in
the first three months mostly due to Digicel and CARE failing to establish an orderly schedule to
undertake field registration procedures and online verification in tandem.
CARE reported such issues occurred due to the extended time required for beneficiary
verification selection and time required by Digicel to perfom online uploading of the information,
37
including registering beneficiaries on site afterwards. According to Transversal, a third party
contracted by Digicel, it required at least 48 hours for registration in a given commune but the
process was rushed in one day due to time constraints.
Schedule issues were coupled with individuals who tried to register several times. According
to the project's director, data as electronic registration was limited in refreshing / reconciling
uploaded data in real time and such entries could only be revealed once data was transferred into
computers and analyzed for duplications. This duplication process was also caused by several
families nominated in different “section communale” due to high rates of changing their residence
address."
Following the first months of the project, the duplication of errors jumped by 397 %.
However, after the first two months of project implementation, the registration process started to
improve significantly with duplication errors dropping by about 13 % within the second month and
decreasing on-goingly for the following months (Figure 3.6).
Additional issues were related to registration process. Digicel agents sometimes mismatched
pin and voucher number. These swapping issues leaded to invalid PIN and vouchers which
cumulated respectively to 3518 and 536 cases for the first two months. over the months these errors
decreased as technicians mastered the technology (Table 3.20)
.
Table 3.20 Transaction errors registered from July to December 2012
Month
Figure 3.6 Duplication errors during program execution
38
Error type July August September October November December
Voucher
disabled
997 332 285 561 1700 30
Voucher
number
invalid
241 295 298 286 92 4
Not enough
money
1725 2150 3089 3286 478 23
PIN invalid 1716 1802 1492 1724 587 15
PIN reset 499 452 410 55 0 0
Mobile phone
not authorized
31 13 11 22 4 0
Total 5209 5044 5585 5934 2861 72
3.3.6 Relevance and Effectiveness of redemption (food delivery)
Redemption transactions efficiency was low during the first two months but then increased
substantially at the fourth month of the program. Redemption is the delivery of food to individuals
registered in the program by a selected shop. Successful redemption is effective when a beneficiary
presents his/her voucher to the merchant and laters receives a message confirming the transaction
(redemption of the voucher amount onto the merchant account) from the system after dialing the
USSD code including beneficiary's voucher number and requesting the beneficiary to enter the PIN
as a confirmation for redemption. The beneficiary then received the preselected basket of food from
the local shops.
As can be noticed in table 3.21, 999 families received their 6th refill in September (meaning
they had completed the cycle of all 6 vouchers), 9051 families in October, and 1562 families in
November. However, as the last refill happened in November, 388 (= 207+181) families missed out
either the 5th or 6th refill. All other families completed the six-month voucher cycle.
Table 3.21. Distribution of the beneficiaries by refill date
Six Months Cycle
Number of
times already
Quantity of
Beneficiaries
August
12.
September
12.
October
12.
November
12.
December
12.
January
12.
39
refilled
Refilled 1 time 181 1st refill 2nd refill 3rd refill 4th refill 5th refill 6th refill
Refilled 2
times 207 2nd refill 3rd refill 4th refill 5th refill 6th refill
Refilled 3
times 1562 3rd refill 4th refill 5th refill 6th refill
Refilled 4
times 9051 4th refill 5th refill 6th refill
Refilled 5
times 999 5th refill 6th refill
Total 12000
Source: Administrative data- extract from table Food voucher analysis September 2012, summary
Significant numbers of errors occurred during the first four months. All combined errors
increased from 5000 near to 6000 from second to fourth month, then dropped considerably by an
average of 20 % for the last two months of the project compared to the peak. Those errors resulted
from an adaptability period due to the use of new technology and difficulties that may have occurred
between beneficiaries and shops during voucher redemption within first four months. While Digicel
mentioned that a Call Center 20211 existed at the beginning of the project, CARE’s staff and
mobilizers claimed it was not really effective for CARE project until the fourth month. It is only at
this period personnel were trained to respond to PIN issues related to the food voucher program.
While some errors were purely technical, some were related to human fraud. That was a sufficient
justification of the next stage of monitoring of the process by CARE staff.
Relevance of redemption process was analyzed in comparison of traditional food assistance
program distribution. During the first three months mobilizers reported that beneficiaries crowded
the lines to redeem their vouchers. The bottleneck of distribution cleared from fourth to sixth month
of voucher redemptions. Moreover, more than 80 % of merchants confirmed that the program was
safe and felt secure during the process (Security and confidentiality, reduction of stigma).
11 202 is client oriented service of Digicel which existed prior of the CARE food voucher program.
40
The beneficiary could pick up food basket without being noticed in the community as being
food assisted. However, low literacy of beneficiaries sometimes compromised their security for PIN
number (Flexibility of redemption, pick up of food).
The beneficiary had a period of 30 days to redeem the value refilled on the voucher each
month. The system facilitated procurement of products through existing market channels alleviating
the burden of coordination from implementing agencies. Overall satisfaction with the system is
analyzed in the section following the monitoring system.
3.3.7 Preferred Form of food assistance
Should the project be implemented again, beneficiaries were asked the form of assistance
they would prefer. The results indicate that they would strongly prefer the basket of the “Kore
L’Avni nou” project (65.27%), while about 17% would prefer more local products, and the same
proportion would opt for more vegetables in the basket (Table 3.22).
Table 3.22 Preference for form of assistance
Product form Percentage
More local Product 16,81
More VegeTables 16,81
As Care did it 65,27
Other forms 1,11
Total 100
The use of the electronic voucher is supported as the best form of intervention by the
majority of the beneficiaries (58.19%). The survey also showed that approximately 16% of the
respondents would prefer cash while about 26% would prefer a combination of voucher and cash.
(See table 3.23).
Table 3.23. Form of intervention
Form Percentage
Voucher 58,19
Cash 15,71
41
Both 26,11
Total 100
3.3.8 Satisfaction with the project by beneficiaries
The beneficiaries were also questioned about their satisfaction with the project. The survey
shows that they were highly satisfied. In fact, more than 72% declared being completely satisfied,
and about 26% appeared to be satisfied. Hence, the proportion of dissatisfied was less than 2%
(Table 3.24).
Table 3.24 Beneficiaries satisfaction with the project
Level of satisfaction Percentage
Completely Satisfied 72,25
Satisfied 25,99
Not Satisfied 0,22
Not satisfied at all 1,54
Total 100
As regards the service provided by the shops, the majority of the beneficiaries appeared to be
satisfied. The survey indicated that about 57% declared being completely satisfied and
approximately 36% indicated that they were satisfied (Table 3.25).
Table 3.25. Satisfaction with service provided by the shops
Level of satisfaction Percentage
Completely Satisfied 56,83
Satisfied 36,12
Somewhat satisfied 4,19
Not Satisfied 1,10
Not satisfied at all 1,76
Total 100
Finally, respondents were asked whether they were informed of the project before its
implementation. The results indicate that 85.33% responded positively to that question, while
42
14.67% had a converse opinion. The first group (positive answer) was also asked whether they
believed they were provided with enough information. More than 46% completely agreed that the
information was enough and nearly 48% agreed it was. (See table 3.26).
Table 3.26. Opinion of beneficiaries on whether they received enough information about the project
Level of agree Percentage
Totally agree 46,21
Agree 47,78
Somewhat agree 0,26
Not agree 4,18
Not agree at all 1,57
Total 100
3.3.9 Satisfaction with the project by shops
The final evaluation also considered the principal challenges faced by the shops related to the
electronic system and interaction with customers (project beneficiaries). Shops owners were asked to express
their level of satisfaction with the project, particularly the impact of this latter on their activities. Details about
these aspects are provided in the following subsections.
All the shops were completely satisfied or satisfied with the project and the security of the electronic
system (Table 3.27). However, one of the shop owner said “he could not say that he was satisfied on his shop
progress”.
Completely Satisfied Satisfied Neither satisfied, nor dissatisfied Not Satisfied
Not satisfied satisfied
at all
Project 100 0 0 0 0
Security of the electronic system 86,4 13,6 0 0 0
Shop progress 81,8 13,6 4,5 0 0
Table 3.27. Shop satisfaction with the project (%):
3.3.10 Program partners and sustainability
Digicel was the main implementing partner of CARE in the electronic food voucher program
in Grande-Anse. Digicel also subcontracted Haitian software developing company, Transversal,
43
and Haitian micro-finance institution, Fonkoze, respectively for developing the electronic payment
platform and for paying shops after redemption. CARE also built relations with community leaders
including local representatives of Haitian Government. This section discusses the role of those
partners, how the project impacted them in a sustainability perspective. The last part also presented
perspectives of the beneficiaries after the project.
3.3.10.1 Role of Digicel and its subcontractors
Digicel is the leading private company providing mobile financial services in Haiti. According to
Dalberg report12, Digicel's mobile money services, TchoTcho mobile13 had a better network to
outreach potential clients through its partnership with Alo Communication and Fonkoze that had 43
branches. Digicel had the following main responsibilities in the project:
 Developing e-voucher platform services
 E-Registration of beneficiaries and shops in the food voucher
 Payment of shops upon voucher redemptions
 Technical assistance/troubleshooting related to the electronic food voucher system
Digicel reported that its existing mobile money platform was not ready for the food voucher
program which lead to contract the Haitian firm Transversal, to develop a tailored platform for the
project. Transversal reported sound capacity building through the development of the electronic
exchange platform called Merchant Pro. As a former grantee of the HI-FIVE project, this experience
improved their capacity to work with non IT client to provide soft services. Through intensive
coaching on programming principles, compliance and information management from CARE Food
Voucher program management, they improved the following features of the basic Merchant Pro
platform:
 Insertion of an accounting module able to be exported to Excel or PDF.
 Insertion of accounting elements for audit purpose: there is distinction between the
demand and approver of a request.
 Redesign of a more friendly interface, more illustrative and less technical.
12 Haiti mobile Money, a point in time case study
13 Mobile money transfer product/service offered by Digicel to Haitian community. E-voucher platform was a separate software
(independent from Tcho-tcho) used for the voucher redemption, as the Central Bank of Haiti has legal limitations and regulations
imposed on the amounts to be cashed in and cashed out through existing Tcho-tcho network, whereas e-voucher platform was flexible
enough to allow unlimited voucher amounts being transferred from beneficiary pre-paid (voucher) account into the merchant account.
44
 Improvement of the interface to be compatible with a low internet debit.
Transversal reported that the project has provided insights and added value to serve non IT
clients. They recruited two more staff to work on this project and estimated now having a better
capacity to respond to client short deadline and providing solutions to humanitarian issues. At the
beginning of the program, Transversal maintained direct contact with the CARE Food Voucher
Coordinator as they were in charge of the all technical aspects related to the platform development.
Once the system became relatively functional and in order to avoid double management and
distorted information flow, information communication (requests of various nature) were channeled
from CARE through Digicel to Transversal. FONKOZE was fully administered by Digicel as their
approved “cash-out agent” supporting merchant payments..
As mentioned above, the role of FONKOZE was to provide liquidity to shops after the
redemption process. FONKOZE received outstanding payment amounts from CARE via Digicel14
and disbursed payment to the shops based on the reconciliation of redeemed voucher amounts
approved by CARE and Digicel (information extracted from the system and verified by hardcopy
receipts managed by each merchant shop). FONKOZE was able to address the bottleneck of
liquidity and proximity with the shops. Shop managers reported 100% satisfaction with the overall
project.
Digicel reported its intention to improve its mobile money platform to serve similar
initiatives in the future. This project provided insights for the scale up these type of services by
decreasing fixed and variable costs. The sustainability of an electronic food voucher scale up is
linked to reliable partners and affordable cost.
3.3.10.2 Sustainability for beneficiaries and shops
The beneficiaries were asked whether they believed they would be able to afford the same
basket of products they received from the project. Only a minority, say 6.9%, believed they would,
while the rest, 93.1%, said no. They were also questioned about their perspective for the future. The
majority of beneficiaries wished the project to be executed again. None of them mentioned personal
or family initiatives to keep on improving their living conditions.
14 Novo Scotia Bank, supporting Digicel financial operations opened two accounts: 1 for CARE to debit the voucher amounts refilled
each month and 1 for Digicel to receive those amounts transferred from CARE account to Digicel account for further cash-out to the
merchants, through Fonkoze branches in rural areas.
45
Concerning shops owners (merchants), all considered increasing shop storage capacity,
selling of local products, and 90.9% of them will keep using receipts as introduced by Food Voucher
program for all future transactions. However, most of them expressed their fear of a probable
decrease in demand after the project. They claimed that the project was more emergency based and
the implementation period was very short for a real change in consumption behavior.
3.3.11 SWOT analysis of the Food Electronic Voucher and scaling up as
implemented
The execution of the project “Kore L’Avni Nou” faced various challenges. Despite these
difficulties, the project exhibited strength, though some weakness aspects have been identified by the
evaluation study (Figure 3.7).
Figure 3.7. SWOT of the electronic food voucher program as implemented by CARE
SWOT analysis highlights are a general assessment of the electronic food voucher as
implemented by CARE in a perspective of scaling up. First, the project contains strong points as
S T R EN G TH S
S e cu rit y a n d co n fid e n tia lit y o f tra n s a c tio n
R e d u c tio n o f stigm a f o r b e n e ficia ry
F le x ib ilit y o f p ic k u p ( d im in u tio n o f c ro w d e d
lin e s)
S tre n g th e n o f tra d itio n a l m a rk e t lin k a g e
C a p a c it y to m o n ito r tra n s a c tio n a n d
b e n e ficia rie s
C o m m u n it y p a rtic ip a tio n in s e le c tio n p ro ce ss (
le g itim a tin g )
T h e p la tf o rm is r e a dy to use for sim ila r
pr oje c t
D e ve lopm e nt of na tiona l firm c a pa c ity to
a c c e ss fina nc ia l se r vic e s
WEAKNESSES
Learning process of the system is about 3 months
Development of technological infrastructure is expensive
This e-voucher technology was tailored to the project's needs
Low level of literacy of beneficiary may compromise
confidentiality of transaction ( PIN)
A third person may redeem the client voucher
O PPO R TU N IT IE S
U s e of sm a rt phone by vendors and introduction of a new
ca rd required only sw iping by bene ficia ry ( identifica tion
of beneficiary by picture , ba r code ).
Introduction of G PS via vendor S m a rtphone to loca te
tra nsa ction a nd m apping m e rchants
S ca le up of the e lectronic food vouche r prog ram ba s ed on
de ve loped infra structure to be used on regular phone.
A s bene ficia rie s dem a nd m ore loca l product , a
com pononent of the project m a y support loca l production
THREATS
Community leaders may use selection of beneficiaries
process for political agenda
Shops beneficiaries of the program may run non
beneficiaries shops out of business.
Electronic Food Voucher of
CARE
46
being to be more effective than traditional food assistance programs. Legitimating target group
selection and ability to monitor this process during the project life was an important feature. In
addition, the electronic system, the security and confidentiality of transactions, the facility of pick up
(redeem) at different places, and the capacity to monitor transactions and beneficiaries were
important aspects that facilitated transactions and avoided (or limited) potential fraud. In case of
similar program implementation in the future, those points should be reinforced.
On the other hand, the project also had some weak points that would need to be improved in
the case of reimplementation. First, respective to a less-one-year project, the learning process of the
targeted system was too ambitious for a new technology system of electronic cash transfer.
Nevertheless this process can be mastered in three months if additional human resources are
committed to the partner side ( Digicel); second, a third person might have redeemed some vouchers
and the confidentiality of the PIN could be compromised due to customers (case of illiteracy for
example) transmitting the PIN to shop owners to make the voucher redemption transaction.
 Regarding threats to project implementation, some community leaders might use and
influence beneficiary selection process for political agendas. Therefore confidence between
implementing agencies and beneficiaries may be compromised. Moreover, as the project selected
shops with higher capital and stock facility, those retailers may create a distortion in local market, as
non-participating retailers may be run out of business.
Finally, implementation of electronic food voucher opens doors for opportunities of
improvement and extension. Similar projects in a sustainability perspective may create link or
partner shops and local farmers to improve local demand elasticity. This strategy should consider
timing of the response of seasonal production, institutional arrangement and economic incentives.
Low literacy of beneficiaries may be counter balanced by longer hands-on training. In addition
extension of e-voucher to regular phone may decrease cost and facilitate financial inclusion (mobile
cash transfer).
47
3.5 Key lessons learned
The study has permitted to learn the following lessons:
 A non-automatic refill helps in the cleaning and verification process.
 Design of an electronic food voucher program should take into account average internet debit in a
given country for access facilitation. Use of an interface user friendly, less technical and more
practical system. System should be able to export data to excel and PDF format.
 Schedule sufficient time between system and field registration for conciliation purpose. Maintain also
constant communication with field team and system operators during the process.
 Selection of beneficiaries by a steering committee may be compromised by personal or political
agenda, therefore a cross-verification or accompaniment of the process by the agency is
recommended.
 An implementing partner of an electronic food voucher program may need one Information
Technology staff to facilitate the process
3.6 Key obstacles preventing the correct execution of the program
During its course, the project faced various types of challenges at different levels. The
following subsections describe the main issues encountered.
3.6.1 Challenges encountered by shops during the project
In terms of challenges, shops owners reported some difficulties with the electronic system
despite that 95.5% claimed they had used phones before the project. In table 4.10, one will notice
that shops had some issues with phone use (27.3%), at least once with payment by phone (50.0%),
with vouchers numbers (45.5%) and PIN (36.4%). 68.2% had received help to solve their problems,
but only 45.2% claimed they were satisfied with this service. According to many shop owners, the
training session received at the start of the project “was good, but insufficient”. (See table 3.28)
T a ble 3.28. D iffic ultie s w ith the e le c tronic vouc he r s ys te m (% ): Y e s N o
P hone us e 27,3 72,7
P a ym e nt by phone 50,0 50,0
V ouc he rs num be rs 45,5 54,5
P IN 36,4 63,6
D id you re c e ive he lp to s olve the m 68,2 31,8
S a tis fie d of this he lp s e rvic e 45,2 54,8
T a ble 3.28. D iffic ultie s w ith the e le c tronic vouc he r s ys te m (% ):
48
The shops related some issues with the beneficiaries, to demand and supply. It seems that
18.2% had some concerns with beneficiaries, mostly with the voucher numbers. 18.2% also affirmed
that sometimes faced shortages of products beneficiaries wished to buy. When that happened, they
offered substitute products.
Sometimes, they had shortages, mostly for local products (27.3%). 50% claimed they
sometimes had difficulties in procuring local products. To explain these difficulties, they mentioned
“the problem of seasonal effect of local products”. (See Table 3.29)
Yes No
Customer 18,2 81,8
Change in customer preference overtime 18,2 81,8
Insuficient stock 27,3 72,7
Over stock 27,3 72,7
Facility in procurement local products 50,0 50,0
Table 3.29. Shop difficulties (% ):
3.6.2 Issues with beneficiaries
According the M&E staff, the level of education of targeted beneficiaries is very low, which
limits their understanding of the full program process. To overcome this, training sessions at the start
of the project should use poster and experimental cases (learning by doing). Illiterate beneficiaries
disclose their secret pin to vendors which increase the probability of fraud. The mobilizers added
that some beneficiaries are better known on their nicknames rather than their official names, which
caused some problems to data collection. Moreover, in some cases, the food was taken on behalf of
the beneficiary by a third person who never took part in the training.
3.6.3 Issues with the community
The mobilizers also reported that they were sometimes victim of intimidation from
community members because they are not selected in the program. This social pressure occurred
most likely in deactivation of the coupon. The Program Coordinator observed that despite
educational information sharing and trainings of the beneficiaries, the project was misperceived by
some communes as a politicized government program. In fact, some individuals associated it with
the electoral process. Furthermore, the lack of confidence of the population in the government put an
obstacle in the relation between CARE and government involvement in the project, though the Civil
49
Protection and Ministry of Social Affairs did not accept to be volunteers or partners in the project;
they rather wanted to be salaried.
3.6.4 Issues with shops
The mobilizers identified the following issues with the shops:
 They lacked business skills to use the receipts.
 Some did transactions without the presence of beneficiaries.
 At the beginning they had issues to compile basket goods and organize logistics.
3.6.5 Issues with the technology
The mobilizers also faced problems with the technology. A transaction might be possible (as
it would be confirmed in a 3 -day delay on the Digicel list), however the basket of goods was not
delivered due to the voucher being non active, invalid or having insufficient credit for the
transaction. Finally, mismatch problems between PIN and selected beneficiaries was also observed.
3.7 Main recommendations
The final evaluation of the project led to the following recommendations:
 Establishing Links between farmers and local shop: The project had injected an important
amount of money through voucher distribution directly into the local economy to boost
spending, especially of the most vulnerable people. As a response to this monetary injection, it
appears that the demand of local products increased during the project execution. However,
according to the shop owners, response time from farmers and seasonal harvesting of local
products were a persistent problem, which complicated the adjustment of the supply to the
demand. However, for a real effect (at mid and long term) on local economy and to ensure the
viability of the project, it is necessary to increase local production by establishing links
between farmers and local shops.
 Reinforcing food insecurity assessment: It would be worthwhile to consider some
anthropometric indicators in selection of the beneficiaries to complement the subjective data.
 Pursue sustainable options: Sustainability of food insecurity improvement is a huge
challenge. It is clear that CARE and partners did what could reasonably be done in the short
term and with limitations. However, the integration of other aspects such as well-run savings
schemes, and expansion of small-scale pilots in nutrition gardens could facilitate some
50
sustainability. Partnering with other agencies should be a top priority for similar programs in
the future.
 Monitoring and Evaluation: Include the commodity market price in the system. As a
reference price for beneficiary shops, the M& E should consider the baseline market price
coupled with inflation. As the biggest retailers are part of the programs, price assessed from
smaller retailers would be bias as reference point. M & E should include gender variable and
introduce more behavioral indicators.
 Period of implementation should be longer: To modify the consumption behavior of
families. Training sessions should be longer, more proactive and practical for appropriation of
the process in the electronic system.
 Strengthen the voucher system by using regular cell phones: The voucher system
presents different signs for effective safeguarding of resources: flexible, secure, efficient in
management, and favorable for customers. However, some fraud cases were registered when
customers had to transmit their PIN to shop owners. Hence, extension of this service to
regular cell phone may be a more effective emergency response tool in the near future.
4. Conclusion
Overall, Kore L’avni Nou has helped to improve the food security status for 12,000 vulnerable
families and reinforced the organizational capacity of 68 local shops. Through various qualitative and
quantitative research methods, this evaluation highlights that beneficiary families greatly benefited from the
program in terms of received food assistance. Management of the electronic food vouchers was very
innovative although being challenging in the first three months.
 Findings showed significant improvement in the living codition of beneficiaries, retailers and local
shops; nevertheless limitations to the scope of the evaluation did not allow confirming all improvements are
relevant only to the project. Beneficiaries were able to redirect spending to other purposes such as child
schooling, livestock assets and other needs.
The electronic food voucher system has a learning curve of three months. The system provided
flexibility for the pickup of food and confidentiality for beneficiaries and all beneficiaries were satisfied with
the system.