Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
78% of Kenya’s population live in rural areas and two-thirds of Kenyans depend on the food
they grow for their livelihoods. Our target areas are the agricultural areas of Lamu, Kilifi with a high
stress level of food insecurity(“FEWS-Kenya,” 2017). This population has been depleting their
household assets and are consuming 1-2 meals per day. The 2017 long rains assessment indicated that
the long rains delayed for about one to four weeks. However, Lamu and Kilifi received 90-125% above
normal. The delayed rains have the potential to disrupt the planting cycle because heavy rains later in the
season have the potential to flood fields. Kenya continues to face lower crop yields. In 2016, maize crop
production was 10% below a five-year average and is expected to decline. Specifically, in Lamu and
Kilifi, this value was about 30%. This has been attributed to short rains and unpredictable weather
patterns. Maize production in the surrounding regions of Uganda and Tanzania have declined as well,
with import volumes as low as 23 percent of the normal. In addition, this has led to reduced market
supplies, which translates to the rise in staple food prices of about 20-40%. Prices have gone as high as
60% of the average market price (“FAO-UnitedNations,” 2017). In May 2017, the government of Kenya
approved the importation of 5.7 million and 6 million of white and yellow maize respectively to help
combat the shortage of food crops and the elevated levels of insecurity (KFSSG, 2017)
Food insecure households have typically intensified coping mechanisms to support food
consumption. These include: skipping meals, reducing meal portion sizes, relying on less preferred
and/or cheaper foods, borrowing food or relying on help from neighbors, friends, and relatives; and
restricting consumption by adults to feed the children (Saaka, Oladele, Larbi, & Hoeschle-Zeledon,
2017). This reflects the importance of supporting farmers through asset microfinance with inputs such as
weather resistant variety of seeds, occasional irrigation support as well as post-harvest storage support
services. Most importantly, farmers do not have access to loans to purchase farm inputs because they
lack collateral. This limits their farming activities, which further worsens their food insecurity status.
The political conflict in the region has led to the inflow of refugees to the pastoral northern provinces of
Turkana and Garissa, a major contributor to the food insecurity level in Kenya(“FEWS-Kenya,” 2017;
WFP, 2016). Due to high demand of food, it is crucial to support local growers to not only support their
households but also allow them to gain income through the sale of staple foods to finance their needs.
Due to the variability in the definition and indicators of food insecurity, we aim to evaluate the
scope of food insecurity that captures food availability, accessibility, consumption, diet diversity as well
as the stress level of households. Our intervention addresses these components in its entirety by
providing assets that improve crop yields, which translates to availability and accessibility, which can
increase income through the sale of foods. Coupled with nutrition education, this will increase the
consumption of a diversified diet and reduce stress, which is a significant symptom of financial
instability, especially in trying to meet the needs of the household. The primary outcome of the
evaluation is to determine the impact of agricultural microfinance on food insecurity. The secondary
outcomes include the increase in crop yield, income, consumption of a diversified diet as well as the
adoption of better farming practices.
Our stakeholders include researchers, program implementers, donors or bank loan, government
agencies and most importantly, the farmers and their communities. It is relevant for government
agencies to be aware of the impact of providing low-interest microfinance programs to farmers to create
a network that will foster economic sustainability for the communities as well as for the country.
Furthermore, with the elevated level of food insecurity projected at 4 million in 2017(“FAO-
UnitedNations,” 2017) due to political instability in surrounding countries, investments in such
programs are essential.
Branches
The aim of agriculture is to harvest desired foods to consume at home to reduce food insecurity
and to sell for income to be able to afford health care, education and to save for unexpected incidents. In
the PIP (Appendix), higher crop yields mean farmers can sell more of their produce to gain income to
purchase animals such as chickens, goats etc. that they can rear to produce meat, eggs and dairy
products. This will increase the consumption of animal produce as well as increase income from the sale
of animal produce such a milk. This also means that if individuals walked to the markets to sell their
produce, they can purchase a bicycle or a motorcycle, which will further increase their assets (“BRAC,”
2013). This feedback mechanism of producing higher crop yields creates a safety net for farmers and
increases overall household income, which has been linked to food insecurity (Gundersen, Kreider, &
Pepper, 2011).
Contextual factors
A factor that influences how effective the intervention could be is the culture of household heads
in the villages. Generally, women are responsible for the nutrition, health, and very often, the farming
needs of the members of the household. However, men tend to be the heads of the household and
therefore, make all the financial decisions. With training and nutrition education, women will be more
likely to grow and purchase fruits and vegetables to diversify their diet (Smith, Ramakrishnan, Ndiaye,
Haddad, & Martorell, 2003). It is possible that men who receive nutrition education could influence their
wives’ decisions, which could also improve food insecurity, but this effect has a greater impact when
women take more of an active role. The empowerment of women has been shown to increase the health
and nutrition status of families, along with improved agricultural outcomes (“Nutrition-Sensitive
Agriculture,” 2016).
Another factor is access to health clinics and other social services. Farming is labor intensive and
requires farmers to be physically capable to engage in these activities. Therefore, they need access to
clinics when illness prevents them from being productive or worse, unable to farm. Depending on the
location and how isolated the community is, this could affect the harvest for the season, which could
mean lower crop yields.
Measurement Bias
We will reduce measurement bias by utilizing evidence-based surveys, such as the Household
Food insecurity access scale (HFIAS) that captures food insecurity status in its entirety in terms of
accessibility, availability, consumption, diet diversity and stress levels. This tool has been used in
developing countries to assess food insecurity levels (Chakona & Shackleton, 2018; M’Kaibi, Steyn,
Ochola, & Du Plessis, 2016; Saaka et al., 2017). By separating treatment and comparison communities,
we can minimize spillover effects that arise due to the sharing of agricultural assets. Furthermore, there
might be difficulties in establishing a temporal relationship because of the duration of our intervention
and measurements. This is because our intervention is occurring during the long rains. Taking
measurements during this period decreases the biases that could result due to different growing
preferences of farmers during the short rains. By separating our education aides from our investigators,
bias will be minimized. Continuous monitoring will allow us to control for unexpected occurrences such
as an outbreak of disease or water contamination that affects some communities and not others. Lastly,
creating buy-in with communities during the previous harvest season will reduce recall bias because
measurements of crop yield can be compared to post-intervention crop yields.
Selection
Farmers who elect to participate could differ in certain characteristics such as less risk-averse,
more motivated or more educated. Farmers who do not participate could also possess similar
characteristics or may be well-off and may not need additional tools. This could overestimate or
underestimate the impact of the intervention. Therefore, by matching treatment and comparison
communities, selection bias is minimized because participants in both groups will share similarities. In
addition, the use difference-in differences approach to measure the food insecurity level before and after
the intervention between treatment and comparison groups mitigates selection bias as well.
Confounding
By limiting participants to females and microfinance seeds to maize, beans and millet, this
decreases the variabilities that gender and food growing preferences add to the evaluation. To reduce
confounding due to different translators for surveys, each matched treatment and control community will
have the same translator.
Part V- Measurements
Indicator Table
Activity Step in PIP Type How we Frequency Indicator used
collected data of
measure
Table 2 gives a brief overview of the planned data measurements. The primary impact for this
intervention is reduced household food insecurity in Lamu and Kilifi in Kenya, while the secondary
impacts are increased crop yield, diet diversity and income for the participants. All the measurements
are used in order to gauge how successful the intervention was to the communities who received the
microfinance as compared to their own baseline and similar comparison communities.
One of the input measurements is the agricultural assets survey, which will be done before and
after the program is implemented. The agricultural assets survey will be used to collect data on how
much assets in the form of agricultural equipment each group took. At baseline, it will also ask them
information on how much income they currently make, then where and what they currently spend their
income on. The survey will be conducted again at the end of the study as well, to asses if their income
increased, and if it changed their purchasing behavior. This will allow us to observe income changes
over the course of the program and allow us to gauge the magnitude of the effect the intervention had on
their income. We are hoping that farmers are spending their money less on food, and ideally more on
education and health.
Agricultural education is the second input measure for the intervention. We will train several
members of the community on agriculture and nutrition to be the education aides for this intervention.
We will provide them with continuous training if need be to ensure their knowledge on these topics. The
educational aides will do bi-weekly meetings at their community center to inform the participants
(fellow community members) of various farming practices, tips and how-to demos and basic nutrition
lessons. The indicator for this measure is an attendance sheet to document how often each participant
comes to the meetings. This is an important input and indicator, because it will tell us when someone is
not attending, how often and how many people total are attending. We can use this to reach out to
people with low attendance to prevent loss to follow-up. One potential threat to internal validity is the
issue of language between the education aides and the people who will be training them, which could
cause an information bias in this intervention. We hope to have fluent people to train the community
members, so when they are education aides they have all the right knowledge to share with their
communities’.
One of the outputs for the program is the loan repayment survey, which will ask questions on
how much money in assets they decided to take and will ask them if they have started to pay back their
loans. If they respond with yes, they we will ask them how much they have paid back. This is something
that will be monitored over time to see when people could start to pay their loans, which is a proxy
measure for when they began to make money from their crops. We also want to collect data on how
much they could start to pay back, we want to see if the effects on their income from the intervention
had an immediate effect for that growing season or not. This will show us if the intervention gave the
participants enough income to support them enough to be able to pay back their loan immediately. The
repayment of the loans by the end of the study does not have to be done, but for those who do, will be
given a fruit tree as a reward. This is used as an incentive to prioritize paying the money back to prevent
the accruing of interest from the microloan.
The agricultural education feedback survey is the other output measurement that we have for our
program. This is a way for the participants to give their input into the program to inquire about what
they want to learn, to provide feedback on how the program is going for them and to specify if they want
the education to change at all or if there are any problems occurring on their farms. When they provide
feedback, the participants will feel more engaged in the process, and will hopefully feel empowered to
learn anything they want at the meetings. The indicator used to measure this is the number of education
meetings attended over the number of crops grown. This is the dose-response measure and we
hypothesize that with more meetings attended, food insecurity will be decreased.
The outcome measure for this program is the food growing survey, this survey is important to
measure how much crops were grown over the growing season. Participants will be provided bags for
the crops. We will ask them to fill the bags with crops, and then write down how many bags full of each
crop they grew, a tally mark of the number of bags will be sufficient. We will weigh the same bags when
they are full of the different crops and translate that into a weight in kilograms.
The Household Food Insecurity Access Scale (HFIAS) is the measure of impact. The
questionnaire asks several questions about food consumption and their feelings about food, trying to
determine if they met criteria for food insecurity. The answers will be translated into a numerical scale
based on presence and severity of the food insecurity. This questionnaire askes basic questions on their
household’s food consumption. It extends to food quantity, diet diversity, and their feelings surrounding
food. Then the total number from their questionnaire will be translated into food secure, mild, moderate
and severe food insecurity. The aim of this intervention is to reduce the amount of moderate and severe
food insecure homes. We want to analyze how much food insecurity was reduced by this intervention.
The main outcome measure will be the amount of food grown. This will show if desired outcome
of more crops grown occurred or not. If the number of crops did not increase, then the impact of reduced
food insecurity may have not happened. The indicator for food grown will be the amount of food grown
after the intervention in kilograms over food grown before the intervention in kilograms. We will
compare the amount grown to their own past seasons, and to the comparison group. We want to know if
the intervention improved the amount of food grown and by how many kilograms. This will be a proxy
measure to see if the intervention was successful for this community, this will see if there was a benefit
seen by the treatment group. The amount of food grown can either be consumed by the household or
taken to market to be sold for extra income. This is an important measure of how successful the
intervention was. However, this cannot be the only measure since if there was some outside factor
(drought, storm, etc.) that prevented the amount of food grown to be increased, the intervention will
seem like a failure when it is not. This could affect the internal validity of the measure, which is why we
also have other measures, like the HFIAS and the agricultural education feedback survey. They are used
as additional measures of success even if the weather patterns are not favorable. The feedback survey
will help us gauge how much they learned, which they can use in the next growing season, and can be
measured as some type of success. Another potential threat to internal validity is if the participants
forget to weigh and tally their crop growth. This will impact the results of this measure. To try to
prevent this, the participants will report their crop growth every meeting.
The main impact measure will be the Household Food Insecurity Access Scale (HFIAS). The
HFIAS is a questionnaire measures the quantity of food consumed, the diversity of food consumed and
feelings and anxiety around food. The indicator for this measure will be the amount of moderate and
severe food insecure homes after the intervention over the amount of moderate and severe food insecure
homes before the intervention (Coates, 2007). This will give us a ratio of how much we reduced
moderate and severe food insecurity in the treatment groups. We will compare this ratio to the
comparison group to see how much it helped reduce food insecurity. We want to make sure our
comparison group has similar food insecurity demographics and similar as possible in all other aspects,
this will help strengthen the internal validity. One issue that could threaten the internally validity is that
after the nutrition education the participants may change their answers to the questionnaire which may
change the result. To combat this potential issue, we will change the wording of the HFIAS to be
explicit and to be as specific as possible, this way the participants will know how to answer the
questions.
Monitoring Information Summary Table
The activities in Table 3 need to be monitored to make sure the program is running smoothly,
and so we can make sure the participants are getting what they need to grow more crops. All the
activities will be monitored at the community center when participants come to the bi-weekly education
meetings. This is the most efficient way to collect data for this program, since the participants will be
coming to one area already. Home visits will not be conducted unless it is necessary.
For this intervention to work properly, we need the participants to be involved in the process.
There will be several benchmark points that will trigger action on our part to ensure maximum
participation. For the agricultural and nutrition education activity, the attendance sheet will inform us of
the number of farmers attending the meetings. If a participant misses more than two meetings in a row, it
will trigger the education aide, who will then reach out to the participant and note the reasons for low
attendance. If attendance overall falls below 80%, it will trigger the investigator to speak with the
participants to see if the education aide is not being helpful, and if any improvements can be done to
increase attendance. Another activity with a benchmark point will be for the agricultural education
feedback. This activity is a survey and will tell the investigators what the participants want to learn, how
the education can improve, how the program is going overall and if there is a problem that we can help
them solve. We want to monitor the progress of the crop growth through this survey. An action will be
triggered when there are improvements that the participants want. Since we want this program to work,
we must cater to unique community’s needs and this will allow us to do so. An example of an action, is
if participants have a pest problem we can provide them with pesticides to help their crop.
Part VI- Timeline
PROJECT TITLE: AGRICULTURAL MICROFINANCE IN KENYA
YEAR: OCT 2018 -FEB 2020
Oct Nov Dec Jan Feb Mar Apr May June July Aug Sep Oct Nov Dec Jan Feb
15-Oct 31-Jan 15-Nov 30-Nov 15-Dec 31-Dec 16-Jan 31-Jan 14-Feb 28-Feb 15-Mar 31-Mar 15-Apr 30-Apr 16-May 31-May 15-Jun 30-Jun 15-Jul 31-Jul 16-Aug 31-Aug 15-Sep 30-Sep 15-Oct 31-Oct 15-Nov 30-Nov 16-Dec 31-Dec 16-Jan 31-Jan 14-Feb 28-Feb
Activities
Phase 1
Training of aides
Buy in with community members
Baseline data collection by investigator
Agricultural land prep activities
Biweekly meeting
Nutrition Education
Delivery of assets within a week
Phase 2
Planting and growing activities
Provision of ongoing support services eg. Irrigation
Harvesting of crops
Repayment of loans
Effective storage processes
Sale of produce
Continues monitoring by investigator
Phase 3
Endline data collection
Incentives
Collective data analysis
The program intervention is for the duration of 16 months. The time frame is due to its
occurrence before the start of the long rains. The intervention is divided in three phases. The first phase
starts with creating buy in with communities, training of agricultural and nutrition aides, baseline data
collection of food insecurity and crop yield, agricultural land prep activities, the start of biweekly
meetings, nutrition education as well as the delivery of assets at least within a week of planting. This
period will be starting from October to December 2018. Biweekly meetings will continue till the end of
the program because loan repayment will be going on during this process. Phase two starts from January
to December 2019 starting with planting activities, provision of ongoing services. The later part of phase
two will include harvesting of staple foods, with the sale of harvests as well as the start of loan
repayment. Continues monitoring by investigators will take place during the biweekly meetings in
addition to the provision of storage services. Phase three includes end line data collection in addition to
the provision of incentives. Collective data analysis will take place within the last two months of the 16-
month process.