Feinstein International Center, Tufts University
PLI Policy Project, Ethiopia
Rapid Review of the Cash‐for‐Work and Natural Resource Management
Components of the RAIN Project
July 2010
Andy Catley and Alison Napier
Contents
SUMMARY...............................................................................................................................................1
Introduction ............................................................................................................................................3
Review process .......................................................................................................................................3
Review workshop....................................................................................................................................4
Rapid overview of cash‐for‐work and natural resource management activities................................4
Asset transfers in Somali pastoralist areas – lessons from other projects.........................................6
Household profiles and simple economic modeling to predict the impact of asset transfers...........7
Linking RAIN to long‐term strategic frameworks for pastoralist areas ..............................................9
NRM activities and impact................................................................................................................ 10
Revisiting the RAIN objectives .......................................................................................................... 11
Findings and recommendations............................................................................................................ 12
Annex 1. Mercy Corps participants at review workshop, Addis Ababa, 28 to 29 July 2010 .................14
Annex 2. Case studies of previous cash distribution and restocking projects......................................15
Annex 3. Trends in ‘high export’ pastoralist areas: linking RAIN strategies to broader development
strategies...............................................................................................................................................19
1
SUMMARY
The project Revitalizing Agricultural/Pastoral Incomes and New Markets (RAIN) is a three‐year
project implemented by Mercy Corps and Save the Children UK (SCUK) in parts of Somali and
Oromiya Regions in Ethiopia. The project aims to protect, build and diversify assets in food insecure
households. The donor is the Office for Foreign Disaster Assistance (OFDA) and the project budget is
US$17 million.  
In July 2010, approximately at the mid‐point of project implementation, Mercy Corps
worked with the Feinstein International Center of Tufts University to review specific elements of the
RAIN project viz. cash‐for‐work (CFW) and related natural resource management (NRM) activities,
and explore opportunities for reshaping project strategies and activities to achieve greater impact.
The review process included the collection and review of case studies from Somali pastoralist areas,
focusing on asset transfer approaches such as cash distributions during drought, and restocking after
drought. Specific areas of interest in these previous approaches were the levels of asset transfer
relative to the livelihoods impact, and the time period needed to achieve impact.  
Complementary to the review of case studies, simple economic modeling was used to
predict how different levels of one‐off cash transfers would impact on different types of destitute
and poor agropastoral and pastoral households. Spreadsheets with the model were provided to
Mercy Corps for further adaptation and to assist revision of project strategies.  
The overall RAIN project objectives were reviewed, albeit briefly, together with options for
assessing the impact of the NRM activities.   
Key findings and recommendations
Project design and strategies
• At the overall level of project design, general implementation strategies, and M&E needs, Mercy
Corps staff had already started to review and reshape specific RAIN activities and question the
likely impact of CFW and NRM activities. However, given the complexity of RAIN and the project
budget, a radical reworking and clarification of project objectives is needed. The current project
design needs to be clarified using sub‐objectives and made SMART viz. Specific, Measurable,
Achievable, Realistic and Time‐bound. Revised activities should fit clearly under, and directly
contribute to the sub‐objectives. At present the RAIN project document lacks a clear causal
pathway, from activities to objectives to impact, and seems not to fully differentiate strategies
or activities according to the characteristics of different households, in different areas. There
seemed to be common agreement in the workshop with Mercy Corps that these changes were
needed.  
• In its current form, RAIN provides one‐off cash transfers which are likely to be very short‐lived in
terms of impact, as the cash is probably used mainly to meet immediate food needs. Therefore,
while RAIN might be contributing to a typical (but unstated) humanitarian/food security
objective, impact on the more development‐orientated objectives of asset protection, asset
building and livelihoods diversification is unlikely. In terms of asset protection, this objective has
to be tailored to the pre‐existing assets of households and does not apply easily to households
which have no productive assets to protect.   
• At the level of CFW activities, the review showed a need to revise these activities if asset
building objectives are to be achieved within the timeframe of the project. Relative to RAIN,
previous and reasonably successful asset transfer approaches in Somali areas have channeled far
greater resources to fewer households, or, have expected impact to last for only two months or
less. Simple economic modeling should help Mercy Corps to reshape CFW approaches, and tailor
different approaches to different types of destitute and/or poor agropastoral and pastoral
households. These approaches need to be clearly defined and justified. There was wide variation
2
in the level of cash transfers within project areas.  RAIN is currently has a very complex design,
whereas the difficult operational context of Somali Region calls for simplicity and focus.   
• For the NRM‐related activities, approaches such as pond and gully rehabilitation, check dam
construction, terracing and soil bunds have been used in Somali and parts of Oromiya regions for
many years. For example, these approaches were widely used by the South East Rangelands
Project in Somali Region. For RAIN NRM activities, potential impacts are currently outlined at
community level and are not specific to destitute or poor households i.e. the most vulnerable
groups. Questions to look at include: How, specifically, will destitute or poor households benefit
from these structures? Given the repeated need for repair or rehabilitation, what local
management systems can be supported to ensure the long‐term maintenance of the structures?
Which structures are susceptible to private ownership with possible exclusion of poorer users? Is
private ownership necessarily a bad thing? Similar questions apply to bush clearing (Prosopis
control), given the rapid reinvasion of cleared areas. Impact assessment of these activities is
likely to be hindered by the good rainfall in 2010 and the likelihood that RAIN‐related water
points or grazing areas might not be used if there is ample ground water available or good
grazing elsewhere.  
• The review did not look at issues such as the effectiveness or relevance of CFW compared with,
for example, direct cash distributions (without work) or food distributions. While different
approaches are likely to result in different impacts and acceptance within communities, there
are also questions about the organizational costs and efficiencies of transfer s, including
operational and transaction costs. For example, in the case of RAIN the CFW approach requires
agencies to plan, support and monitor the tied NRM activities, and at a cost. At some point, an
evaluation of RAIN should assess the ‘delivery costs’ of different asset transfer approaches.
Long‐term, the administrative and operational costs of international NGOs are likely to far
exceed those of some other service providers.  
• RAIN is but one project among many, and its objectives and strategies should fit within a
broader, long‐term Mercy Corps framework for development in pastoralist areas. The review
touched on some of the issues. As a ‘high livestock export’ area, Somali Region and neighboring
Somali areas are subject to long‐term but gradual commercialization. This is reflected in the
robust nature of the export trade and simultaneously, rising levels of destitution. As
commercialization advances, herds are likely to increase and poorer herders will struggle to stay
in pastoralism. If correct, these trends have major implications for Somali areas and
development policy, and the strategic directions of RAIN.       
Implications of change
• Mercy Corps staff expressed concerns over donor expectations, and pressure to maximize the
number of project beneficiaries in RAIN. Despite these concerns, as a humanitarian donor OFDA
may want to revisit the concept of achieving development objectives via asset transfers which
are spread widely but thinly, and which are one‐off transfers. Should OFDA support a radical
reshaping of the CFW component of RAIN, it may be acceptable to use different approaches in
similar livelihoods zones, and compare impacts. This might involve a continuation of the CFW in
some areas as is, but piloting more intensive transfers to fewer households in other areas.  
• By using a CFW approach in RAIN, Mercy Corps has become engaged in debates and
negotiations in and around the Productive Safety Net Programme, especially on wage rates. In
its current form in pastoralist areas, the CFW approach is probably contributing to a short‐term
food security needs for some households but having limited impact on asset growth. This in itself
might be a lesson to feed into the PSNP and wage rate discussions. In the face of the next shock
or drought, households are still likely to deplete assets rapidly and require assistance. The
economic  modeling tool developed as part of this review might be further developed and used
with government partners to predict the impacts of cash and/or food transfers, and the levels
and duration of transfer needed to achieve meaningful asset building objectives.   
3
Introduction
The project Revitalizing Agricultural/Pastoral Incomes and New Markets (RAIN) is a three‐year
project implemented by Mercy Corps and Save the Children UK (SCUK) in parts of Somali and
Oromiya Regions in Ethiopia. The project aims to protect, build and diversify assets in food insecure
households. The donor is the Office for Foreign Disaster Assistance (OFDA) and the project budget is
US$17 million.   
At around the mid‐point of the implementation of RAIN, in May 2010, Mercy Corps
approached the Feinstein International Center (Tufts) to seek support for an impact assessment of
specific RAIN activities viz. cash‐for‐work (CFW) and related natural resource management (NRM)
activities of the project. When deciding if an intervention should be assessed and when, experience
from the Pastoralist Livelihoods Initiative (PLI) in Ethiopia from 2005 has shown that impact
assessment is not always appropriate. For example, a desktop review of project design and
implementation strategies might indicate that limited impact would be expected and if so, a
reshaping of project design can be more useful than impact assessment1
.  
With this experience in mind, Tufts reviewed the RAIN project proposal to OFDA and
following discussion with Mercy Corps, agreed to support a technical review of the project aims and
strategies, and with a focus on the CFW and NRM components. In part this focus was also influenced
by wider debates around safety net programs in Ethiopia, and the role of cash and/or food transfers
as a means to reduce vulnerability to livelihoods crises.   
Review process
The overall purpose of the review was to examine what level of income/assets are required to
enable households to maintain or improve their livelihoods, especially in the face of pressures such
as drought.  
Activity   Type of information and analysis
Desk review (Tufts)
Workshop/seminar
with RAIN staff
• Review lessons from other assessments in Somali Region related to the Productive Safety Net
Programme (PSNP) and/or CFW (e.g. Save the Children US, World Bank), other cash transfer
programs  
• Review lessons from restocking programmes e.g. the herd growth rates and core herd sizes
required to maintain or grow herds
• Review of trends in pastoralist livestock holdings  
• Review of existing CFW and wage rates (public and private sectors) in target areas
• Review of food economy baseline data for RAIN areas
Based on the above, develop a number of ‘real’ profiles or case studies, of different household
economies. This could include profiles of a pastoralist household in a rural area, a female‐headed
household, an agro‐pastoralist household, a destitute or ‘drop out’ household etc. and will include
households likely to have been targeted by CFW/ PSNP programs.
• Based on the different household profiles, current asset status and lessons learned, with
Mercy Corps staff develop scenarios to illustrate and compare the potential livelihoods impact
of different development options over time.
• Variables could include: different types and levels of inputs (CFW vs. restocking); livelihood
shocks (price rises, drought); different timeframes etc.  
• Using the scenarios, identify a) the inputs/ support needed to protect current financial assets
and b) the different economic options (including those targeted by RAIN) for improving/
diversifying livelihoods in the short to medium term.  
• Discussion on the extent to which RAIN objectives and strategies fit with higher‐level donor or
Mercy Corps strategies in pastoralist areas of Ethiopia.  
                                                           
1 For example, a review of the Mercy Corps activities in Afar Region under PLI was conducted in late 2006 in preference to
an impact assessment.   
4
Potential next steps
This could lead into a discussion of the current RAIN cash for work strategies and possible options
for strengthening the approach.
Based on the discussions during the workshop but could include field work to develop additional
case stories (for example, of households that have successfully recovered/ diversified) and/or
identifying key questions for examining the potential impact of RAIN’s NRM work, for example in
relation to different household economies/ scenarios.
It was expected that the process, especially the workshop, would include an element of
capacity building of Mercy Corps staff in the use of simple scenario analysis.  
The workshop was held at the Mercy Corps office in Addis Ababa 28th to 29th July 2010, with
Mercy Corps staff (Annex 1). There were no participants from SCUK.  
Some of the key background documents which were reviewed by Tufts and used during the
workshop were as follows:
• Evaluations of previous asset transfer/asset‐building projects in Somali areas – these were
summarized and used as case studies during the workshop (see Annex 2).
• Food economy baselines for the Harshin Degahbur East Pastoral Livelihood Zone and the
Jijiga Agropastoral Livelihood Zone, published by SCUK and Disaster Prevention and
Preparedness Agency; the reference year was 2004‐5.  
The review had initially intended to examine PSNP reviews in pastoralist areas, particularly those
supported by SCUS and the World Bank. However, these documents were not available at the time
of the RAIN review.
Review workshop
Rapid overview of cash‐for‐work and natural resource management activities
Mercy Corps presented a rapid overview of CFW and related NRM activities in three project areas viz
East Hararghe, Jijiga and Degabur.  
East Hararghe office
Beneficiaries ‐ the emergency CFW targeted destitute households; the NRM CFW targeted poor
households.
Table 1. Cash transfers in East Hararghe
Type of CFW Total transfer (ETB) Number of beneficiaries Average total cash transfer (ETB)
per household1
Emergency CFW
NRM CFW
957,220
557,243
803 (607 female, 803 male)
1010 (284 female, 726 male)
696 (~US$26)
1372 (~US$51) 1
Based on two workers receiving cash per household; EB 13.4 = US$1.
Table 2. NRM activities in East Hararghe
      
Type Number Volume
Ponds (water pans)
- New construction
- Rehabilitation
Gully rehabilitation, including reshaping, check dam
construction, plantation and cut‐off drains
2
9
4
1,600m3
2,500‐4,000m3
37,088m3
Expected outcomes from emergency CFW:
- Cash necessary for beneficiaries to purchase vital food supplies (i.e. no need to sell assets)
5
- Population stayed in area
- In good/favorable position to benefit from seed fail activities  
- Improved access to water
Expected outcomes from NRM CFW
- No need to sell assets for food
- Asset building: purchase of livestock (small ruminants)  
- Improved access to water
- Improved environmental services
- Livelihood diversification through income generating groups
Jijiga office
Table 3. Cash‐for‐work by activity in agro‐pastoral areas
Location and type of work Women Men Total amount paid
(EB)
Average total payment
(EB) per person
Gursum; wage rate EB 20/day:
‐ shallow well construction
‐ hillside terracing (10km)
‐ pond rehabilitation (1800m3
)
Kebribayah; wage rate EB 25/day:
- hillside terracing (33km)
‐ stone check dam (4318m3
)
‐ soil bund (27km)
20
30
40
116
66
89
30
280
180
234
260
284
50,000
50,000
72,000
212,500
431,800
101,250
1,000
161
327
607
1,324
271
Note – targeted poor agro‐pastoral households with no cereal in store.  
Table 4. Cash‐for‐work by activity in pastoral areas
Location and type of work Women Men Total amount paid
(EB)
Average total payment
(EB) per person
Aware ‐ wage rate EB 30/day:
- stone check dams (500m3
)
‐ soil bund (36km)
‐ Propsopis clearing (100ha)
Gashamo ‐ wage rate EB 30/day:
‐ stone check dam (2000m3
)
‐ water diversion channel (40km)
‐ soil bund (80km)
‐ pond rehabilitation (91,200m3
)
5
10
5
10
5
6
0
75
110
55
90
85
97
50
60,000
162,000
99,000
240,000
180,000
198,000
72,000
750
1,350
1,650
240
2,000
1,900
1,400
Note – targeted poor pastoral households with < 50 sheep and goats. Maximum 2 workers per households, though this was
unusual; most households had only 1 worker in the project.  
Degahabour office
Table 5. Cash‐for‐work by activity
Kebele Activity Quantity Number of
workers
Total amount paid
(EB)
Average total
payment (EB) per
person  
Ararso
Lafgaloli
Obale
Pond rehabilitation
Pond rehabilitation (incomplete)
Gully rehabilitation
5,250m
1,250m
1,000m
350
182
100
262,500  
67,125  
50,000  
750
368
500
- The project sought the most vulnerable households, but left the selection to a “CFW
committee”; this led to occasional mis‐selection, which led to disagreements over wage
rates (EB 25/day).
6
- The activities aimed to deal with water shortages, particularly in Obale.
- Preliminary findings at field level were that cash was used to buy food; for example, a
household of 7 people might buy 50kg wheat, lasting 1 week and costing EB 100; with an
average weekly CFW wage of EB 150, wheat could be purchased to last about 1.5 weeks.
Asset transfers in Somali pastoralist areas – lessons from other projects
This session was based on Mercy Corps staff reviewing a series of short case studies from previous
restocking projects and cash distributions in Somali pastoralist areas. The full case studies are shown
in Annex 2. An important part of the review was to examine the levels of asset transfer relative to
the reported impacts and benefits in these projects (summarized in Table 6), and then compare to
the levels of asset transfer in pastoralist areas of RAIN (Table 4).  
Table 6. Previous evaluations and impact assessments, Somali areas2
SC Isiolo, Kenya SC Fik, Ethiopia NORDA, Kenya Horn Relief, Sool
Context Drought‐related IDP rehabilitation Drought‐related Drought‐related
Type of transfer Cash Livestock, food, other Livestock Cash
Key objectives Long‐term herd
reconstitution;
diversified livelihoods
Return of IDPs to
pastoralist livelihood
Resumption of
pastoralism
Short‐term
emergency
assistance, for 1.5 to
2 months
Value of asset
transfer
US$490/hh US$321/hh plus food
ration
~US$450 US$50/hh
Frequency of asset
transfer
One‐off One‐off One‐off One‐off
Achievement of
objectives
Partial at 7 months
after cash input; up to
further 2 years
needed to build herds
75% hhs resumed
pastoralism 2.5 years
after asset transfer
Most hhs attained
minimum herd size
1.5. years after
restocking
Yes, but impact of
assistance lasted only
1 month
Table 4 indicates that the average total cash transfer per household in pastoralist areas of RAIN
was US$122 (EB 1677). A comparison of this figure with Table 6 indicated that the cash transfers
provided by RAIN:
- would last a family of around 7 people about 2 to 3 months if all cash was converted to
food;  
- would probably have minimal impact on asset building if the strategy was to ensure that
pastoralists acquired a minimum herd. The restocking case studies used transfers of US$321
to US$490 per household, with the lower figure supplemented with a food ration.  
These findings indicated a need to redesign the asset building approach for pastoralist
households and specifically, to provide a higher level of cash transfer to a smaller number of target
                                                           
2 Sources for Table 6:
Acacia Consultants Ltd. (2004). Evaluation of cash relief programme implemented by Horn Relief, commissioned by NOVIB/Oxfam
Netherlands. Acacia Consultants, Nairobi
Ali, D., Toure, F. and Kiewied, T. (2005). Cash relief in a contested area: Lessons from Somalia. Humanitarian Practice Network Paper 50,
Overseas Development Institute, London
Croucher, M. et al. (2006). Initial Impact Assessment of the Livelihoods Programme in Merti and Sericho. Save the Children Canada,
Nairobi
O’Donnell, M. (2007). Cash‐based Emergency Livelihood Recovery Programme, Isiolo District, Kenya – Project Evaluation draft report, by
Michael O’Donnell, Save the Children, Nairobi.
Lotira, R. (2004). Rebuilding herds by re‐enforcing gargar/irb among the Somali pastoralists of Kenya: evaluation of experimental
restocking program in Wajir and Mandera Districts of Kenya. African Union/Interafrican Bureau for Animal Resources, Nairobi
and Feinstein International Center, Tufts University, Nairobi
Wekesa, M. (2005). Terminal evaluation of the restocking/rehabilitation programme for the internally displaced persons (IDPs) in Fik Zone
of the Somali Region of Ethiopia. Save the Children UK, Addis Ababa and Acacia Consultants, Nairobi
7
households. Plus, given the remaining time in RAIN for implementation it was noticeable that even
with relatively high asset transfers it would take around two to three years to rebuild pastoral herds.   
Household profiles and simple economic modeling to predict the impact of asset transfers
Another way to examine the likely impacts of cash transfers to pastoralist and agro‐pastoralist
households is to develop simple economic models which show patterns of asset growth over time,
and in the face of different pressures on livelihoods such as drought. Simple modeling can be used as
a tool to answer questions such as:  
o For different levels of asset transfer, how long will it take a destitute pastoralist household
to develop sufficient assets to resume pastoralism as the main means of livelihood?
o For different levels of asset transfer, how long will it take a poor pastoralist household to
‘move upwards’ into a medium wealth group category?
A first step in the modeling process was to characterize different types of agropastoral and
pastoral household by asset ownership, bearing in mind that the objectives of RAIN focused on asset
protection, asset building and asset diversification.  The profiles in Table 7 were suggested by Mercy
Corps staff for RAIN project areas.
Table 7. Households assets in RAIN project areas
Even before using these profiles for modeling, they pointed to some basic questions in terms of the
objectives of RAIN. For example,  
• For different types of household, which type of assets – specifically‐ is RAIN trying to protect
and why?
• For destitution households, does the objective of asset protection apply if these households
have so few assets to start with?
For the purpose of illustration, an initial model was developed for a Somali household of 7
people, being 2 adults and 5 children. The models used livestock herd dynamics data, data on human
nutritional requirements and energy content of foods, and data from household economy surveys
on the proportional contribution of different foods to household diets. The models focused on asset
growth and separate models were developed for two different types of household viz:  
Destitute agropastoral
household, East Hararghe
Poor agropastoral
household, East Hararghe
Poor agropastoral
household, Gursum
Poor pastoral household
• No grain stores
• No livestock
• Basic household items
e.g. cooking utensils,
bedding
• Clothes
• Qat in fields but crop  
not suitable for sale
• Land
• Exhausted social
support
• Children probably not
going to school
• 2‐3 sheep or goats
• Perhaps 1 oxen
• No cash
• Small seed reserve
• Basic household items
e.g. cooking utensils,
bedding
• Clothes
• Qat in fields but crop
not suitable for sale
• Land
• Better ability to labor
• Better social support
networks
• About 5 sheep or goats
• About 2 cattle
• No cereals stored
• No cash
• Qat in fields but crop
not suitable for sale
• Basic household items
e.g. cooking utensils,
bedding
• Clothes
• Less than 50 sheep or
goats
• Cattle ownership rare
• Basic household items
e.g. cooking utensils,
bedding
• Clothes
8
• Destitute pastoralist households ‐ with starting pre‐project livestock assets of 6 sheep/goats.  
• Poor pastoralist households – with starting pre‐project livestock assets of 40 sheep/goats, 3
cattle and 3 camels.  
For each model, two common assumptions were made:
• All scenarios assumed a drought every four years which depleted livestock assets by 30%.
• All scenarios assumed food aid receipts of 50kg maize/hh/year.
Five scenarios were developed as follows:
• For destitute pastoralist households three models were developed:
o one model assuming no external asset transfer;
o one model assuming cash transfer of US$75 (EB 1030);
o one model assuming livestock transfer valued at US$350 (EB 4795).
• For poor pastoralist households two models were developed:
o one model assuming no external asset transfer;  
o one model assuming cash transfer of US$75 (EB 1030).
The initial modeling results are shown in Table 8.  
Table 8. Results from simple economic modeling of Somali pastoralist households  
Type of household and input Scenario result Notes
Destitute pastoralist, pre‐
project livestock of 6
sheep/goats.
1. No external asset transfer
2. One‐off cash transfer of
US$75
3. One‐off cash transfer of
US$350 (or livestock
equivalent)  
Poor pastoralist households,
pre‐project livestock of 40
sheep/goats, 3 cattle and 3
camels
1. No external asset transfer
2. One‐off cash transfer of
US$75  
After 5 years this household will have
livestock assets of ~25 sheep and goats
only, well below the minimum herd size for
pastoralism.
After 5 years this household will have
livestock assets of only ~39 sheep and
goats, still below the minimum herd size for
pastoralism.
After 2 years livestock assets are ~ 60
sheep and goats, and 2 cattle; this is above
minimum herd size.
After 3 years, livestock assets reach
‘middle’ wealth group category
After 3 years, livestock assets reach
‘middle’ wealth group category
These households will rely heavily on social
transfers and external assistance e.g. food
aid.  
These households will rely heavily on social
transfers and external assistance e.g. food
aid.
As above, assumes 50kg maize as food aid
per year. Transfer is equivalent to that
used in more successful restocking
programs.
As above, assumes 50kg maize as food aid
per year. Removing this food aid has
limited impact on asset growth.
As above, assumes 50kg maize as food aid
per year. Removing this food aid has
limited impact on asset growth.
9
Note that:
• For destitute pastoralist households, initial transfers of value approximately US$350 plus
food are required if they are to acquire a minimum herd size within a typical program time
period.   
• For poor households which already own a minimum herd, low cash transfers of around
US$75 have limited impact on asset growth.
This kind of information reflects the lessons and challenges from restocking programs in
pastoralist areas, from the 1970s viz. for a given budget is it better to channel relatively high
livestock (or cash) transfers to a relatively small number of households, or, provide small transfers to
a large number of households?
Excel spreadsheets with the basic model were provided to Mercy Corps staff, who started to
adapt the models for agropastoral households and test the impact of different levels of asset
transfer. Together with the case studies, simple modeling seems to be a useful approach to help
staff think through project strategies and for given levels of asset transfer, likely impacts on different
types of household within the project timeframe.  
Linking RAIN to long‐term strategic frameworks for pastoralist areas  
For agencies supporting various projects in pastoralist areas, there are coordination and strategic
issues around the ways in which project outputs combine (or not) to achieve long‐term strategic
development objectives. In PLI, there has been initial discussion around the use of the livelihoods
framework as a tool for developing strategic plans with NGO partners, and related capacity‐building
options. In April 2010, Mercy Corps supported a combined livelihoods‐conflict analysis in Shinile
Zone of Somali Region, leading to a proposed strategic framework for that area3
. Recent research in
Somali areas also shows the trends towards commercialized pastoralism and the implications for
development policies.4 In summary, commercialization is associated with a gradual transfer of
livestock from poorer to richer households, making it increasingly difficult for poorer herders with
fewer animals, to stay in the system.  
With these issues in mind, a review of RAIN asset protection/building/ diversification
objectives and strategies could take note of the main trends which affect livelihoods in Somali
Region, and the concept of ‘moving up and moving out.’ Those people moving up are able to acquire
and maintain large herds, are characterized as middle‐wealth or wealthy, and increasingly engage in
commercial export markets. Those people moving out are gradually selling out to bigger herd
owners. For these people the livelihoods options will include some potential for employment and
service provision around the livestock sector. For many others, the options are limited in pastoralist
areas, and this partly explains increasing destitution and increasing migration. One line of argument
is that safety nets and similar approaches, with relatively low levels of asset transfer, may encourage
poorer people to remain in pastoralists areas, when long‐term, non‐livestock based economic
opportunities are limited.        
A brief presentation was made on these issues (Annex 3) followed by discussion with Mercy
Corps staff. Issues included the RAIN focus in productive assets, implying economic/financial assets,
whereas as key strategy for diversification in pastoralist areas is improving education i.e. a form of
human capital.  
                                                           
3 Catley, A. and Iyasu, A. (2010), Moving Up or Moving Out? A rapid livelihoods and conflict analysis in Mieso‐
Mulu Woreda, Shinile Zone, Somali Region, Ethiopia. Mercy Corps and Feinstein International Center, Addis
Ababa (in press) 4 Aklilu, Y. and Catley, A. (2009). Livestock Exports from the Horn of Africa: An analysis of benefits by
pastoralist wealth group and policy implications. Feinstein International Center, Tufts University, Addis Ababa
https://wikis.uit.tufts.edu/confluence/display/FIC/Livestock+Exports+from+the+Horn+of+Africa  
10
NRM activities and impact
This part of the workshop used a selection of three RAIN NRM activities, with discussion on
proposed indicators for assessing the impact of these activities. The activities and proposed
indicators from Mercy Crops were as follows:
Pond rehabilitation Gulley rehabilitation Bush (Prosopis clearing)
Increased access to water
- Distance to closest water
point
- Length of time during which
water can be accessed
- Number of animals using
water point
- Change in expenditure for
water
Improved sanitation – number of
people using clean water
Cash transfers
Reduced rate of land loss
- Increased arable land for
agriculture and grazing
- Increased potential in food
self‐sufficiency or income
generated
Reduced risk of flood damage (if
overall water management considered
in design)
Output/outcome:
Improved rangeland condition
Improved access to grazing
Impact:
Improved household food security
(contribute to)
Could be measured by:
- Number of households who
have got access to improved
grazing
- Body condition of animals
   
Some general issues arising from these indicators were as follows:
Causal pathways and defining where impact should happen
Who, specifically, is supposed to benefit? RAIN objectives refer to ‘food insecure’ households, but
this could be more clearly defined. Food economy surveys define poor, and in some cases, destitute
households, each with different levels of assets. Is it these households who might benefit from the
NRM inputs and if so, how? Do the potential benefits vary depending on whether a household is
poor or destitute? In other words, for each type of target household in a given area, what is the
causal pathway between the repair or creation of a structure, and specific livelihoods impacts at
household level?  
Meaningful indicators
Quantitative indicators need both a numerator and denominator. An indicator such as ‘1000 people
access clean water’ is potentially misleading if the number of people who need clean water is, for
example, 20,000. A better indicator would be ‘50% of households within 5km of clean water source’.  
Targets
Targets are useful. If an indicator is intended to measure a change, the degree of change can be
specified e.g. ‘clean water availability improved from 8 months/years to 11 months/year’. The target
here is 11 months of availability.  
Language
Avoid terms in indicators which are difficult to define or measure e.g. what do the following terms
means: ‘improved grazing’, ‘increased potential’, ‘reduced risk’?
Methods
Many indicators for NRM activities can be defined and measured using participatory mapping.
Where map have already been produced e.g. for planning purposes, the same maps may contain
useful baseline information.  
Management and access
An important aspect of NRM activities is the long‐term management of new or repaired structures.
This is reflected in the apparent need for external support to repair or rehabilitate some structures.
There are also issues around who, specifically, can access water, improved grazing or land for
11
cultivation. Can these resources be used by poorer households and if not, why not? What are the
risks of appropriation of resources by wealthier or more influential people? If follows that NRM
activities should include a set of indicators around local management capacities and systems.     
     
What if it rains?
If structures are intended to improve access to water or grazing during dry periods or drought, they
might not be used if years of good rainfall. If so, it becomes more difficult to assess impact. This is
relevant to RAIN because 2010 in particular is likely to be a year of high rainfall in much of Somali
Region.   
In part, useful indicators can flow naturally from well‐defined project objectives and/or sub‐
objectives. The project objectives are discussed in the following section.
Revisiting the RAIN objectives
A project document should describe a logical flow of activities to objectives. Tools such as logframes
and/or the use of SMART objectives5 help to ensure that the technical and programming logic of a
project is clear. For project working in different areas with different socio‐economic characteristics
(or livelihoods zones), the project document should also specify how activities and strategies vary by
area. It is also important to define target groups, and to do this as specifically as possible.  
The RAIN project document has two objectives:
1. Food insecure households protect their agricultural and pastoral productive asset base and
prepare themselves for participation in more profitable markets (budget US$5.1 million).
2. Food insecure households increase and diversify their asset base via immediate economic
opportunities and the development of high impact agriculture and non‐agricultural markets
that spur private sector investment and local economic growth (budget US$11.9 million).
Neither of these objectives are SMART. However, it seems that in OFDA proposals objectives are
sometimes broadly written because a change of objectives requires OFDA approval, which takes
time. It follows that one approach to clarifying RAIN would to develop SMART sub‐objectives.
Options include:
- Replace the term ‘food insecure’ with a more specific description of target households in
terms of the wealth of the households (e.g. poor, destitute), the livelihood system (e.g.
agropastoral, pastoral), and the physical location. This approach immediately helps to focus
attention on specific types of household in specific areas; also see the household profiles in
Table 7.
- Sub‐objectives under Objective 1,  
o Consider the meanings and measurement of terms like ‘prepare’, ‘participation’ and
‘more profitable’. What, specifically, do these terms mean and how might they be
measured? If terms are vague or difficult to measure, they should be replaced with
terms which are easier to define and measure;
o Be careful to relate ‘asset protection’ to household type; destitute households may
have no productive assets to protect and if so, does the objective apply to these
households?
- Sub‐objectives under Objective 2,
o Increases in, and diversification of assets need to be defined by targets and
quantified measures, again, by household type and area;
                                                           
5 SMART – Specific, Measurable, Achievable, Realistic, Time‐bound.
12
o Is the development of high impact agriculture achievable during the three years of
RAIN and if so where, specifically? – it seems not to apply to pastoralist areas;
o Are non‐agricultural markets relevant to pastoralist areas, given the major role of
livestock in pastoral economies?  
o What is the meaning of ‘…spur private sector investment’ and how will this be
measured? – private sector investment in what, specifically?
o How is ‘local economic growth defined’ – on an area or household basis? If an area
basis is used, it implies inclusion of all households, including the wealthy. If so, a
project which targets the wealthy may achieve local economic growth while making
the poor poorer.    
Having reviewed the RAIN objectives against SMART criteria, there seemed to be common
agreement that RAIN needs substantial, perhaps even courageous, reshaping and focusing if asset
protection and asset building objectives are to be reached. Even then, these objectives are likely to
be achieved in far fewer households than described in the project document. So far, RAIN staff have
worked to adjust some of the activities. However, higher‐level changes at the Objective/Sub‐
objective level are needed to rationalize the project. Revised activities can then evolve from the
revised objectives.
Findings and recommendations
Project design and strategies
• At the overall level of project design, general implementation strategies, and M&E needs, Mercy
Corps staff had already started to review and reshape specific RAIN activities and question the
likely impact of CFW and NRM activities. However, given the complexity of RAIN and the project
budget, a radical reworking and clarification of project objectives is needed. Sub‐objectives are
needed which are SMART viz. Specific, Measurable, Achievable, Realistic and Time‐bound.
Revised activities should fit clearly under, and directly contribute to the objectives. At present
the RAIN project document lacks a clear causal pathway, from activities to objectives to impact,
and seems not to fully differentiate strategies or activities according to the characteristics of
different households, in different areas. There seemed to be common agreement in the
workshop with Mercy Corps that these changes were needed.  
• In its current form, RAIN provides one‐off cash transfers which are likely to be very short‐lived in
terms of impact, as the cash is probably used mainly to meet immediate food needs. Therefore,
while RAIN might be contributing to a typical (but unstated) humanitarian/food security
objective, impact on the more development‐orientated objectives of asset protection, asset
building and livelihoods diversification is unlikely. In terms of asset protection, this objective has
to be tailored to the pre‐existing assets of households and does not apply easily to households
which have no productive assets to protect.   
• At the level of CFW activities, the review showed a need to revise these activities if asset
building objectives are to be achieved within the timeframe of the project. Relative to RAIN, past
and reasonably successful asset transfer approaches in Somali areas have channeled far greater
resources to fewer households, or, have expected impact to last for only two months or less.
Simple economic modeling should help Mercy Corps to reshape CFW approaches, and tailor
different approaches to different types of destitute and/or poor agropastoral and pastoral
households. These approaches need to be clearly defined and justified. There was wide variation
in the level of cash transfers within project areas.  RAIN is currently has a very complex design,
whereas the difficult operational context of Somali Region calls for simplicity and focus.
• For the NRM‐related activities, approaches such as pond and gully rehabilitation, check dam
construction, terracing and soil bunds have been used in Somali and parts of Oromiya regions for
many years. For example, these approaches were widely used by the South East Rangelands
13
Project in Somali Region. For RAIN NRM activities, potential impacts are currently outlined at
community level and are not specific to destitute or poor households i.e. the most vulnerable
groups. Questions to look at include: How, specifically, will destitute or poor households benefit
from these structures? Given the repeated need for repair or rehabilitation, what local
management systems can be supported to ensure the long‐term maintenance of the structures?
Which structures are susceptible to private ownership with possible exclusion of poorer users? Is
private ownership necessarily a bad thing? Similar questions apply to bush clearing (Prosopis
control), given the rapid reinvasion of cleared areas. Impact assessment of these activities is
likely to be hindered by the good rainfall in 2010 and the likelihood that RAIN‐related water
points or grazing areas might not be used if there is ample ground water available or good
grazing elsewhere.  
• The review did not look at issues such as the effectiveness or relevance of CFW compared with,
for example, direct cash distributions (without work) or food distributions. While different
approaches are likely to result in different impacts and acceptance within communities, there
are also questions about the organizational costs and efficiencies of transfer s, including
operational and transaction costs. For example, in the case of RAIN the CFW approach requires
agencies to plan, support and monitor the tied NRM activities, and at a cost. At some point, an
evaluation of RAIN should assess the ‘delivery costs’ of different asset transfer approaches.
Long‐term, the administrative and operational costs of international NGOs are likely to far
exceed those of some other service providers.  
• RAIN is but one project among many, and its objectives and strategies should fit within a
broader, long‐term Mercy Corps framework for development in pastoralist areas. The review
touched on some of the issues. As a ‘high livestock export’ area, Somali region and neighboring
Somali areas are subject to long‐term but gradual commercialization. This is reflected in the
robust nature of the export trade and simultaneously, rising levels of destitution. As
commercialization advances, herds are likely to increase and poorer herders will struggle to stay
in pastoralism. If correct, these trends have major implications for Somali areas and
development policy, and the strategic directions of RAIN.       
Implications of change
• Mercy Corps staff expressed concerns over donor expectations, and pressure to maximize the
number of project beneficiaries in RAIN. Despite these concerns, as a humanitarian donor OFDA
may want to revisit the concept of achieving development objectives via asset transfers which
are spread widely but thinly, and which are one‐off transfers. Should OFDA support a radical
reshaping of the CFW component of RAIN, it may be acceptable to use different approaches in
similar livelihoods zones, and compare impacts. This might involve a continuation of the CFW in
some areas as is, but piloting more intensive transfers to fewer households in other areas.  
• By using a CFW approach in RAIN, Mercy Corps has become engaged in debates and negotiations
in and around the Productive Safety Net Programme, especially on wage rates. In its current
form in pastoralist areas, the CFW approach is probably contributing to a short‐term food
security needs for some households but having limited impact on asset growth. This in itself
might be a lesson to feed into the PSNP and wage rate discussions. In the face of the next shock
or drought, households are still likely to deplete assets rapidly and require assistance. The
economic  modeling tool developed as part of this review might be further developed and used
with government partners to predict the impacts of cash and/or food transfers, and the levels
and duration of transfer needed to achieve meaningful asset building objectives.   
       
14
Annex 1. Mercy Corps participants at review workshop, Addis Ababa, 28 to 29 July 2010
• Talew Dheresa
• Emma Proud
• Jeff Shannon
• Abdinasir Mohamed
• Ahmed Osman
• Tsegaye Hagos
• Fasil Demeke
• Mark Dwyer
• Retta Aklilu
Facilitator – Andy Catley, Tufts University
  
15
Annex 2. Case studies of previous cash distribution and restocking projects
Isiolo, Kenya Case Study for Review of the Mercy Corps RAIN Project 
July 2010 
• Read the case study below, from a pastoralist area of northern Kenya. 
• Note the similarities between the objectives of the Kenya project and RAIN in terms 
of building assets and alternative/diversified livelihoods. 
• Note the level of cash transfer per pastoralist household, at ~US$490/hh; this is 
equivalent to ~Eth birr 6713 using exchange rates in July 2010. 
• Note the modest food security impacts after seven months. 
• What questions might this raise in terms of the design of RAIN and options for 
reshaping strategies and design? 
Livestock Emergency Guidelines and Standards 
http://www.livestock-emergency.net 
Case Study 9.1: Herd reconstitution using cash transfers, Kenya 
Isiolo District in Kenya’s Eastern Province suffered from a severe drought in 2005 that resulted in high livestock
deaths and elevated acute malnutrition rates among infants. Following improved long rains in April/May 2006,
Save the Children Canada provided 750 households in 22 communities with a one‐off cash transfer of KSh
30,000 (approx. US$490; this is equivalent to ~Eth birr 6713 using exchange rate in July 2010). The cash was
intended to assist families to reconstitute their herds with animals of their choice or to invest in alternative
productive uses, and also to have some cash to meet pressing immediate needs.  
On average, livestock prices at local markets did not change significantly as a result of the cash
distribution, although sellers did attempt to charge exorbitant prices because of the sudden increase in
demand. Beneficiaries adopted a variety of methods for dealing with this attempted inflation, including
purchasing as groups with a representative, travelling to more distant markets, and delaying their purchases.  
An evaluation was carried out seven months after the distribution, which found that recipients
appreciated the cash‐based intervention because it gave them the choice to purchase the specific animals of
their choice and exert more quality control than is possible with in‐kind restocking. It also allowed recipients to
spend some of the cash on other needs.  In total, 85% of the cash was spent on livestock – mainly goats, sheep
and cattle, with some donkeys. The remaining 15% was split between items such as shelter construction,
investing in business/petty trade, debt repayments, veterinary care, healthcare, education and food. Children’s
attendance at school, especially for girls and at the secondary level, has increased for the recipients compared
to non‐recipients.  
The programme targeted only 11% of all households, and hence clearly did not reach all of those in
need. However, this was linked to the availability of funding, and it was agreed that it was better to provide
larger amounts of cash to a smaller number of people than to spread the available money more thinly across
all those in need.  
Seven months after the cash distribution, the impact on food security has been modest. Recipients
have improved the diversity of their diet, especially because of increased access to milk; however their reliance
on food aid has not been significantly reduced. Based on herd growth in the first 5‐7 months (+ 3% for cattle, +
16% for goats and + 25% for sheep), it has been estimated that herds should be large enough to ensure food
security within 2 years, which is substantially faster than if there had been no intervention.  However, the final
impact of the programme will only be clear in the longer‐term and in particular during the next drought when
the beneficiary households’ resilience will be put to the test.  
Sources: Cash‐based Emergency Livelihood Recovery Programme, Isiolo District, Kenya – Project Evaluation
draft report, by Michael O’Donnell, Save the Children, May 2007
Initial Impact Assessment of the Livelihoods Programme in Merti and Sericho by  Matthew Croucher, Victor
Karanja,  Rukia Wako, Abdikadir Dokata and Jillo Dima. Save the Children, 2006
 
16
Fik Case Study for Review of the Mercy Corps RAIN Project 
July 2010 
• Read the case study below, from Fik in the Somali Region of Ethiopia. 
• Note the similarities between the objectives of the Fik project and RAIN in terms of 
building assets. 
• In the case study note the level of asset transfer per household. 
• Note the recommendations of the evaluation in terms of increasing the value of the 
initial asset transfer to 50-70 sheep and goats, while also providing food for at least 
six months. 
• What questions might this raise in terms of the design of RAIN and options for 
reshaping strategies and design? 
_________________________________________________________________________ 
Case Study: Restocking in Fik, Somali Region, Ethiopia 
• A one‐year project, implemented between March 2002 and August 2003.
• Objective of re‐integrating 500 vulnerable IDPs in Fik Zone into their home communities through
improved capacity to build assets for a pastoral life.  
• The project was funded and implemented by SCUK, in collaboration with other stakeholders, notably
government partners, CRDA and UNICEF.  
• The package comprised 30 small stock, 1 donkey, 1 plastic sheet, 1 blanket, 250kg of maize and
provision of animal health services through the training of 10 animal health workers. The cost of the
project was Eth birr 2.2 million, or ~Eth birr 4400 (US$ 321) per household.
• The project was implemented in 11 sites in Fik Zone, based on a detailed participatory scoping study
undertaken in early 2002.  
Some results:
• The contribution of livestock to food sources rose from 2% before restocking to 40% after restocking.
At the same time food relief dropped from 50% before restocking to 7% after restocking. This means
that livestock and by products had effectively replaced food relief as a source of food, signifying
increased self‐reliance and improved household food security.  
• Of those beneficiaries interviewed, a total of 75% (38 out of 51) of restocked IDP households claim to
have moved out of IDP camps and gone back to a pastoral lifestyle, one and a half years after the end
of the project (that is 2.5 years after the project started).
• Most restocked households pointed out that “food aid is very important but it does not reproduce like
livestock. Livestock is a better food source because it builds up during the good years and gives the
family some dignity and respect from other community members”.
Lessons:
‘From this project, the lesson is that the restocking package was useful but too small and should comprise of
the following:  
• At least 50‐70 sheep and goats in preferred proportions and right age for immediate breeding
• An adequate food ration comprising of cereals, oil, sugar and tea leaves for a period of at least six
months
• Provision of adequate veterinary services by trained CAHWs and basic animal health knowledge and
skills to be given to beneficiary households.’  
Note that increasing the package to 50‐70 sheep and goats/hh would increase the value of the initial asset
transfer to ~Eth birr 7500/hh (US$ 547).  
Source
Wekesa, M. (2005). Terminal evaluation of the restocking/rehabilitation programme for the internally
displaced persons (IDPs) in Fik Zone of the Somali Region of Ethiopia. Save the Children UK, Addis Ababa and
Acacia Consultants, Nairobi
17
Wajir-Mandera Irb Case Study for Review of the Mercy Corps RAIN Project 
July 2010 
• Read the case study below, from Somali areas of northern Kenya. 
• In the case study note the level of asset transfer per household. 
• What questions might this raise in terms of the design of RAIN and options for 
reshaping strategies and design? 
_________________________________________________________________________ 
Case Study: Using traditional Somali restocking to assist asset building in Kenya 
• Implemented by NORDA after drought; aimed to between March 2002 and August 2003.
• 400 poor pastoralist households targeted.
• Beneficiaries already owned 5 to 10 sheep and goats, and 2‐3 cattle or camels; the project provided
an additional 15 sheep and goats and 1 donkey per household; through the traditional irb system, the
community provided an additional 5 sheep and goats per household i.e. the total asset transfer was
20 shoats and 1 donkey per household.  
Some results:
An evaluation was conducted 1.5 years after households had been restocked.
• Most of the surveyed beneficiary families had attained the “minimum survival” flock/herd size of 40
shoats.
• Diseases were a major threat to herd growth. Common diseases included Fever (Qano/Tuya) and CCPP
(Ferefekle/Riwein).
• The restocking significantly contributed to reduction in dependence on other sources of food from
84.1% to 57.9% and increased access to food through livestock from 15.9% to 42.1%.
• Dependence on other sources of income dropped from 91.3% to 60.3% and this was compensated by
increased reliance on income from sale of livestock and livestock products from 8.7% to 39.7%.
• The surveyed beneficiary families had substantially re‐established themselves as pastoralists albeit
from a small restocking package.
• The beneficiary families were not only afforded the chance to return to pastoral lifestyle but also the
opportunity to re‐unite and strengthen social ties with family members they separated from during
drought.
• The willingness of the community to contribute for and support beneficiary families was widely and
hugely unreserved.’
Sources of food
before and
after restocking
Source
Lotira, R. (2004). Rebuilding herds by re‐enforcing gargar/irb among the Somali pastoralists of Kenya:
evaluation of experimental restocking program in Wajir and Mandera Districts of Kenya. African
Union/Interafrican Bureau for Animal Resources, Nairobi and Feinstein International Center, Tufts University,
Nairobi
0
5
10
15
20
25
30
35
40
45
50
Selling
livestock
Meat Milk Relief Remittances Wild products,
Casual
labour, $
business
Others
Sources of food
Population (%)
Before
After
18
Sool Cash Relief Case Study for Review of the Mercy Corps RAIN Project 
July 2010 
• Read the case study below, from areas neighbouring Somali Region of Ethiopia. 
• In the case study note the level of asset transfer per household and compare to the 
average incomes received by beneficiary households under RAIN. 
• Note the duration of the impact/benefits provided by the Horn Relief programme. 
• What questions might this raise in terms of the design of RAIN and options for 
reshaping strategies and design? 
_________________________________________________________________________ 
Case Study: Cash relief program in Sool and Sanaag regions, Somalia/land 
• Implemented by Horn Relief during drought, December 2003 to March 2004.
• 13,380 pastoralist households targeted; cash grant of US$50 per household.
• Food was available in local markets, but limited purchasing power of households.
• The objective was ‘Increase the purchasing power of vulnerable populations to meet food and
essential non‐food needs with the provision of a one‐time grant of US$50.’ This grant is equivalent to
Eth birr 685 using exchange rates in July 2010.
• The impact of the transfer was expected to be time‐limited, to about 1.5 to 2 months.
• Targeting was based on criteria for excluding households, being those with more than 60 sheep and
goats, and ownership of assets such as berkads or water tanks.  
Some evaluation results:
• 97% of beneficiaries met the targeting criteria.
• Main uses of the cash grant were meeting needs in food and water, health and debt repayment.
• Although the benefits were viewed as substantial in terms of food security, these benefits were
estimated to have lasted only 1 month – a larger cash grant would have been needed to extend the
benefits.
• Impacts on local markets and businesses were positive, but again, short‐lived.
Sources
Acacia Consultants Ltd. (2004). Evaluation of cash relief programme implemented by Horn Relief,
commissioned by NOVIB/Oxfam Netherlands. Acacia Consultants, Nairobi
Ali, D., Toure, F. and Kiewied, T. (2005). Cash relief in a contested area: Lessons from Somalia. Humanitarian
Practice Network Paper 50, Overseas Development Institute, London.
  
19
Annex 3. Trends in ‘high export’ pastoralist areas: linking RAIN strategies to broader
development strategies
  
20