Delivery of Social Protection Programmes in Kenya

By Clemensia Mwiti and Nupur Kukrety

Clemensia Mwiti is a Humanitarian Support Professional (HSP) in Emergency Food Security and Livelihoods.

Nupur Kukrety is the Social Protection and Food Security Advisor with Oxfam GB based in Oxford, UK.

This article would not have been possible without the invaluable support provided by Claudie Meyers, the Livelihood and Food Security Coordinator for Oxfam GB Kenya Programme, in gathering all the relevant documentation and making appointments with the relevant offices. We are equally grateful to the Hunger Safety Net Project (HSNP) field staff in Turkana, especially Rebecca Lomalia, the Programme Officer who coordinated the field visits and took us through the practical implementation processes, highlighting challenges and lessons learnt.

The article has also benefited from the insights of the HSNP Secretariat Coordinator, Sammy Keter, based in the Office of the President who gave an elaborate description of the project, highlighting gaps in Phase 1 and recommending areas for strengthening before commencement of Phase 2.

It was however, not possible to circulate the draft to all the people mentioned, and any inaccuracies or misinterpretation are unintentional and remain the sole responsibility of the authors.

This article describes Oxfam's early experiences in the Hunger Safety Net Programme, specifically piloting methodologies for delivery of social protection programmes. Based on a field visit to Turkana District, Kenya, it may not necessarily reflect the experience in other districts of the programme.

Oxfam GB, in collaboration with Care Kenya and Save the Children UK (SC UK), are implementing a Hunger Safety Net Project (HSNP) in Kenya, funded by the Department for International Development (DFID) at a cost of £87.2 million. The project is one of three projects that DFID is supporting, costing £122.6m over a 10 year period and aimed at cutting extreme poverty, hunger and vulnerability in Kenya. The HSNP will be implemented in two phases. Phase 1 (£28.1 million) covers 60,000 beneficiaries and will run from August 2007 - July 2011. This pilot phase will develop an evidence base for effective targeting and delivery approaches for cash transfers and will inform the Phase 2 scale-up. Phase 2 (£59.1 million) will run from August 2011 to July 2017 and will contribute to the rollout of a national social protection strategy and an implementation framework, covering 1.5 million Kenyans.

The social protection framework has been designed to enable national scale-up at an estimated cost of Ksh 7.6 billion (£56.3m) for the social transfers, which the Government of Kenya (GoK) is expected to factor into its budgetary commitments. This figure represents less than 2% of total GoK expenditure and approximately 0.5% of national income. It is considered sustainable based on experiences from other low income countries like Lesotho and Nepal that have sustained similar programmes.

Launched in April 2008¹, Phase 1 covers four districts of the arid and semi-arid lands (ASAL) (Turkana, Marsabit, Wajir and Mandera) where poverty levels range between 72% and 92% according to government statistics. This phase targets the hardcore poor population defined either as those among 20% of the poorest nationally or those falling within the poorest quintile.

The project administrative consortium of Oxfam GB, Care Kenya and SC UK were awarded the targeting consultancy by DFID because of their Hunger Reduction activities in the ASAL area of Kenya. The Consortium is supporting the Phase 1 inception of the project by contributing their extensive cash transfer and social protection programme experiences, and are well placed to test various targeting methodologies in support of informing decision-making on appropriate methodologies to be rolled out during the Phase 2 scale-up.

Programme Design

The primary goal of the HSNP is to reduce extreme poverty in Kenya through guaranteed cash transfers to chronically food-insecure households. As cash transfers on a national scale are new and untested in Kenya, the HSNP is designed both to pilot methodologies that can effectively target the poor and transfer small amount of cash efficiently to a large number of people. The programme also seeks to determine whether such transfers can have an impact on poverty and hunger reduction. Phase I has a research component that seeks to demonstrate a costeffective way of transferring cash to the poorest in some of the most challenging environments in Kenya that have limited infrastructure, high insecurity and highly mobile communities.

The programme design comprises five independent components implemented by different Management Consultants (MC):

• Administration component
  Led by Oxfam GB in partnership with CARE International and SC UK, this aims to identify, design and implement appropriate administrative processes for the effective delivery of the HSNP. This includes developing a manual on how to implement the identified targeting methodologies, as well as designing and implementing systems for registration, targeting, enrolment and case management in co-ordination with the other four components.
• Social Protection Rights (SPR) component
  This is headed by HelpAge International with the key mandate of protecting the rights of beneficiaries and other stake holders within the community.
• Management Information Systems component
  This concerns the set up of an information management system for information sharing across all the other components through recording, storage and reporting of key household information and generation of reports that facilitate effective management of the programme.
• Payments component
  Managed by Equity Bank, this component is responsible for disbursing cash to beneficiaries that are enrolled into the programme, based on the different targeting methodologies. The component is piloting the use of biometric data in the profiling of the beneficiaries through use of fingerprints, which is being used in Kenya for the first time, to issue payment smart cards (see box 1). The distribution of the payment smartcard is the last contact between the Administration component field team and the beneficiaries in the framework of the targeting process. The role of the Administration component translates into case management, upon delivery of the smartcards.
• Monitoring and Evaluation (M&E) component
  Headed by Oxford Policy Management (OPM)², this is responsible for project evaluation based on a randomised controlled evaluation methodology. The M&E Managing Consultant will randomly assign 24 treatment locations and another 24 as control locations. The targeted households that receive the cash transfers will form the treatment group for M&E purposes. The targeted households within the 24 control sub-locations who do not receive the cash transfers will receive a 'goody' bag (of an equivalent value of one payment, i.e. 2150Ksh) immediately after being assigned to the control group and are guaranteed enrolment to cash transfers after a maximum 2 years period.

The overall project design and implementation processes allow for the adaptation of the programme design when encountering unforeseen bottlenecks or design flaws. Figure 1 illustrates the central operating systems under the HSNP. This is designed to ensure a seamless integration of administration (registration) and payment systems.

Targeting methodologies

Oxfam GB is operational in Turkana and Wajir districts, while CARE Kenya works in Marsabit district and SC UK in Mandera district. Several targeting methodologies were screened and three were selected for piloting on the basis of effectiveness, acceptability, replicability, scalability and ethical concerns.

Community based targeting
Community based targeting uses a criterion that is developed jointly with the community, based on the definition of poorest households for each community. The criterion, therefore, differ from one community to another and from district to district. The following are the key targeting steps in community based targeting:

Step 1: The secretariat first approves the sub locations for targeting in the districts. This is followed by a one day meeting by the project staff for the relevant leaders to share information about the project at the district level. Invitees include local community leaders, government representatives from the affected sub location and interested politicians.. During the leaders meeting, staff also collect information regarding key sub-location statistics like population figures, migration patterns, number of villages and distances. This information is later triangulated with secondary data and that from other sources to be used for targeting purposes. The plan for targeting is also shared and agreed on with the leaders and the leaders are mandated to pass the information to the community.

Step 2: Staff visit the community and carry out social mapping of the sub locations.

Step 3: One week after the leaders' meeting, a public community meeting (baraza) is organized and is facilitated by the administration and SPR components. During this meeting the project objectives are explained to the beneficiaries and method of targeting shared. Review of the meeting is done the following day with representatives from the government to triangulate any information collected and to resolve any basic statistical differences, where they exist.

Step 4: Five day meetings are organized in each village with the community to agree on the targeting criteria. The first three days are devoted to the administration component to discuss targeting. The last two days are led by the SPR component to inform people about their rights and processes, as well as support systems that have been put in place.

Step 5: The selection process now commences, starting with the mandatory criteria. The primary selection criteria are based on level of poverty. If there are four criteria agreed by the village, these are listed in order of priority.

Step 6: Three days are given before the community is invited to another one day meeting when selection and registration of potential beneficiaries based on the criteria are done. A transparent random sampling is used, that gives everybody a chance to say who qualifies. The process starts with the spin of a pen, and the person it points to is asked to list five households that qualify for registration, based on the agreed criteria. After the naming of the households, the households are discussed by the community members until a consensus is reached on their status. If any is considered not to meet the criteria, then the household is consensually dropped. Someone from among the identified households is then asked to name a further five households that qualify based on the criteria and the process continues till the quota is met. Selection of beneficiaries is based on criteria set in step 5 and follows a chronological order. After criterion one is exhausted, then selection moves to criterion two and so on until the quota for the village is exhausted. Sometimes it might not be necessary to go through all the criteria, if the quota is met.

Step 7: The names of target beneficiaries are now registered by the staff of the administration component partner along with names of a primary recipient and two secondary recipients³. The names of the beneficiaries are pinned in agreed public places in each village (such as church, mosque, community centres, shops, schools etc) for public scrutiny and validation by the entire village. Copies of the potential beneficiary lists are also given to the chiefs and the rights committees. The list remains on display for validation for 14 days and within this period, any complaints and grievances regarding the potential beneficiaries are registered with the SPR committee and forwarded to the administration component partner.

Figure 1: A flowchart for major systems expected for the Hunger Safety Net programme

 

Step 8: After the 14 day's validation period, a meeting is facilitated by the rights committee. The meeting provides a transparent forum for the community members to provide comments on the tentative beneficiary list and propose amendments to address exclusion errors, if any. The community feedback is taken into consideration and the administration field team takes action on the proposed amendments. A final list of agreed beneficiaries is produced with the community.

Step 9: After registration processes are over, a day is given for enrolment. During enrolment, photos and fingerprints of the beneficiaries and recipient and other vital information (e.g. ID numbers if not given during registration) are taken.

Step 10: The enrolment list is then forwarded to the head office of the administration consultant for screening and verification before onward transmission to the secretariat. The HSNP secretariat coordinator then screens and passes this list to the Management Consultant for the payment component to prepare smartcards in the name of recipients. The smartcards are made in the name of the recipient to allow collection of cash on behalf of the beneficiary.

Step 11: Equity bank receives the beneficiary lists, prepares smartcards, which are then sent to the nearest Equity Bank branch together with the list of recipients. Oxfam GB staff collect the cards and distribute them to beneficiaries.

Step12: Case management takes place where complaints are brought to the administration component and are addressed and beneficiary lists are continually updated not to affect payments.

The whole process of targeting and final enrolment takes one and a half months.

Social Pension targeting (Pure Categorical)

The social pension targeting criteria is based on an 'on-demand approach', with all those aged above 55 years⁴ coming forward to register. Unlike community based targeting, this criterion targets individuals, not households. All persons who can prove that they are over 55 years are included in the programme.

Steps 1-3: These are the same as for community based targeting. After the five day training, each sub location is split into three spots that will be used for registration. A day is then communicated when the registration will take place and potential beneficiaries report to these sites for registration. Allowance is made for mobile registration through home visits for those who cannot come to the registration spots.

Step 4: On the registration day, all those who qualify, report to the registration sites. The registration takes place under a tree and the process involves the potential beneficiary providing proof that they are over 55 years. The process is very tedious given that many people do not have identification cards and even those who possess ID cards may have inaccurate information on their age. To overcome exclusion errors, a vetting committee is formed that comprises village chief and some active members of the community who are above 55 years. Key events such as migration patterns, disasters, elections and published records on such events are used. The rights committees set up for the project and other local leaders play a vital role in validation and age verification. However, all individuals who come to the registration desk are registered, whether they meet the eligibility criteria or not. Being registered does not mean that the individual will receive the HSNP benefit, but this will allow follow-up. It is therefore important to clarify to the individual that being registered does not grant them the right to HSNP benefit.

People who have no identification and claim to be 55 years old have to prove that they are 10 years older, i.e. they have to prove they are 65 years before they can be included in the programme. This mechanism is put in place because the vetting committees are unable to establish the exact dates, so events in a span of a 10 years slot help in ascertaining the age. Other facts that are verified include residency in the area.

In order to limit exclusion errors, an extra day is allocated after the registration to accommodate those who have received the information late or have not been able to reach the venue. Meanwhile, staff visit the houses of those with limited mobility in order to register and enroll them

Step 5: After the registration is finalized, the names are read out to the community. The list is then posted in the agreed public places and copies given to the appropriate chiefs and the rights committees. During the 14 days validation period, all grievances are passed to the rights committee. During this period, registration can still continue for those that were unable to make it to the registration sites.

Step 6: After the 14 days, the enrolment date is given and the same steps as the community based targeting (Steps 9 to 12) are followed for enrolment.

Dependency Ratio Targeting (Categorical)

A dependency ratio targeting methodology depends on how many household members are able to work and provide income and how many are dependent on those who work. It gives a good approximation of structural poverty. The process involves registration of all households in the target sub location with a list of all the household members and their ages. In this method the dependency ratio is calculated using a given formula⁵ for each household and the ratio is compared to a threshold that captures the ability of one household to meet its basic needs.

Steps 1-3 are similar to community based targeting, however the criteria for targeting is not revealed until all the households have been registered with details of all household members, to avoid cheating.

 

Step 4: After the five day training, registration is conducted for all beneficiaries at agreed village centres. The registration involves household registration and recording the names and age of all individual household members, with documentation of any form of disability that affects any member of the household. After the registration is complete, the names are posted in all the agreed strategic places and the validation and grievances process follow the same route as the other targeting methodologies.

Step 5: The targeting criteria are only given after the 14 days validation period, based on the list of the entire household members. The dependency ratio is calculated and used to provide provisional beneficiary numbers. Discussions on family residency, composition, ages and disability of household members takes place after the registration process is complete and lists pinned up. At this point, amendments can be made to the list if some households are missing or some information is misrepresented.

Step 6: After validation of the lists, the formula for calculating dependency ratio is shared and enrolment of beneficiaries begins. After the formula has been revealed, there can be no changes to the list. At the time of enrolment the families bring age verification documents. The targeting agency normally calculates the number of beneficiaries during the window period of 14 days to save time.

Step 7: After enrolment of the beneficiaries, the process follows the same steps as the community based targeting (Steps 10-12 above).

Programme Implementation Framework

The core targeting procedures for all the three methodologies are similar and involve seven core steps⁶, which can be summarised into three major categories as illustrated in Box 2.

The process starts at the sub-locational level, which is the smallest targeting unit in each district. The target sub-locations are agreed between the administration and SPR components and are passed on to the Secretariat for approval before identification of beneficiaries can start. In all the methodologies, the entire population living in the selected sub-location is screened. Any individual or household that meets the specific targeting criteria is eligible for inclusion in the project. In the case of Community-Based Targeting, the number of beneficiaries is defined before the field team arrives in the field, based on the rationale for the targeting methodology and on demographic data available for the targeted sub-location using the quota system. For the Social Pension and Dependency Ratio targeting methodologies, there is no quota for the registration and enrolment of beneficiaries.

The targeting process is conducted in 'waves' and each wave produces a number of enrolled beneficiaries that are forwarded to the secretariat for approval. The original project design was staggered into 12 waves, and each wave was to produce an average of 5,000 beneficiaries. However, due to delays in project start-up, this has been changed to 8 waves, with each wave producing approximately 8,000 beneficiaries. The field teams are expected to produce a list of beneficiaries at the end of every month till 2010.

Early lessons from implementing the administration component

Achieving the right balance between transparency and complexity

The entire targeting process from the first meeting with government staff to the communication of the list of beneficiaries to the HSNP secretariat can typically take up to 2 months. An additional 2 month period is taken up processing the payment, so that it takes approximately 4 months for beneficiaries to receive their first payment.

It can also be a challenge to implement such intensive processes in areas where populations are not sedentary, leading to exclusion errors. Moreover, in areas with high poverty headcount, it can be frustrating for many poor households to go through the targeting process and then realise that they do not qualify for the benefit. Most importantly, governments in low income countries may not possess the capacity for implementing such processes in a transparent and efficient way. For the purpose of scale-up to cover large populations (especially by governments), it is important to consider simpler yet robust processes. For example, universal transfers are recommended by social protection experts for areas with high a poverty headcount. In such cases, the cost of limiting inclusion errors can be more than the additional cost required to cover all households.

The myth about categorical targeting

In the social protection debates, categorical targeting (particularly that based on age criteria) is presented as a simple and efficient way of reaching out to the most needy. The experience in HSNP, however, does not support this argument - this programme is operating in areas where people do not possess any documentation, such as ID cards, that contain information about their age. The challenge therefore is to ascertain age of potential beneficiaries through community based processes. In HSNP, vetting committees have to be formed for each sub-location and these committee members trace a set of events and peer group criteria to ascertain eligibility. This can take as much, if not more, time than the other targeting approaches being piloted in this programme. Categorical targeting may not, therefore, be the most appropriate and easiest option in all contexts.

Avoiding exclusion errors

In HSNP, a conscious decision has been taken to concentrate on avoiding exclusion errors rather than focus on reducing inclusion errors. For example, visiting houses of potential beneficiaries who are physically unable to come to the registration site and providing a window period to allow for inclusion of beneficiaries who have been missed during registration.

Quality of staff

The design of HSNP is such that staff require multiple technical and social skills. Staff working on the administration component require a high degree of sensitivity and patience, as they deal with illiterate and innumerate populations on a daily basis. Every effort has been made to select staff with these skills and to support the development of such skills where necessary.

Importance of a clear organisation structure

In its current form, there is considerable lack of clarity on the structure and the roles played by different players and this has impacted the functioning of the administration component. There have been difficulties with the relationship between the secretariat and the steering committee. The HSNP coordinator has lacked sufficient powers to make appropriate decisions and this partially contributed to the delay in starting the programme. This aspect is currently being streamlined by HSNP management.

Do multiple components help?

Having separate components in the project may not be efficient. Some components like administration, rights and M&E can easily be merged together for implementation by one management consultant. This would greatly reduce the transaction costs and enhance the effectiveness of the programme. Reducing the components at the scale-up stage might increase efficiency of the project.

Ownership by national government

Although the project has government endorsement, linkages with government that emerge may require to be strengthened. HSNP is not represented in high level government meetings or in presentations on social protection by government representatives in external fora. If the national government is not proactively involved in the pilot phase of HSNP, it will not be part of learning from the challenges being faced in this phase. This will in turn affect the design and implementation of the phase 2 scale-up in which the government is expected to play a more fundamental role.

Working as part of an NGO in consortium

This has proven useful but has also posed a set of challenges. The biggest challenge has been to ensure compatibility of procedures and financial systems between different agencies.

For further information, contact: Clemensia Mwiti, email: cmwiti@oxfam.org.uk or clemwiti@yahoo.com and Nupur Kukerty, email: Nkukerty@oxfam.org.uk

---

¹The Standard Newspaper Article of 24th April 2009.

²http://www.opml.co.uk/

³The beneficiary and recipients can be different people. The function of the recipient is to collect the money on behalf of the beneficiary.

⁴55 years was the official retirement age for Kenya when the programme was designed, it has since been raised to 6

⁵Dependency ratio =(100 x No. children (1-4 yrs) = No. elderly (55+ yrs) + No. disabled (15-54yrs)) /(Number of people in household)

⁶1. Preparation for targeting exercise 2. Introduction of the HSNP to the leaders and the community 3. Registration of the potential beneficiaries 4.Community validation of potential beneficiary lists 5. Enrolment of beneficiaries and recipients 6. Distribution of smartcards and 7. Case management