80968 KENYA NATIONAL SAFETY NET PROGRAM FOR RESULTS Technical Assessment The Technical Assessment of the Kenya National Safety Net Program for Results was prepared between November 2012 and May 2013 by Judith Sandford with contributions from Ian MacAuslan and support from Sarah Coll-Black, Michael Munavu, Cornelia Tesliuc, and Will Wiseman. The document was edited by Fiona Mackintosh. Material and feedback were gratefully received from colleagues within the Ministry of Gender, Children and Social Development, the National Drought Management Agency and development partners. Citation: World Bank (2013), “Technical Assessment of the Kenya National Safety Net Program for Results”, World Bank, Washington D.C. Table of Contents 1 INTRODUCTION .................................................................................................................................... 1 1.1 CONTEXT .................................................................................................................................................................... 1 1.2 PROGRAMME DESCRIPTION ................................................................................................................................... 2 1.2.1 Government Programme ............................................................................................................................ 2 1.2.2 World Bank Programme ............................................................................................................................. 5 2 STRATEGIC RELEVANCE ..................................................................................................................... 6 2.1 APPROPRIATENESS OF THE INTERVENTION TO NEEDS .................................................................................... 6 2.2 ALIGNMENT WITH THE GOVERNMENT OF KENYA’S PRIORITIES ................................................................... 7 2.3 THE OPPORTUNITIES PRESENTED BY INVESTING IN THE SECTOR ................................................................ 8 3 TECHNICAL SOUNDNESS .................................................................................................................... 9 3.1 SCALE, SCOPE, AND TARGET GROUP ................................................................................................................. 10 3.2 TARGETING, CHANGE MANAGEMENT, AND RECERTIFICATION ................................................................... 11 3.2.1 Targeting ........................................................................................................................................................ 11 3.2.2 Change Management and Recertification ....................................................................................... 18 3.3 MIS AND SINGLE REGISTRY ................................................................................................................................ 20 3.4 PAYMENTS .............................................................................................................................................................. 22 3.4.1 Generosity ....................................................................................................................................................... 23 3.4.2 Payment Mechanism ................................................................................................................................. 25 3.4.3 Frequency, Timeliness, and Predictability ....................................................................................... 27 3.5 ACCOUNTABILITY AND GRIEVANCE PROCEDURES ......................................................................................... 29 3.6 BUILDING THE CAPACITY OF THE NSNP TO SCALE UP SUPPORT IN RESPONSE TO SHOCKS ................ 31 3.7 COMPLEMENTARY ACTIONS AND CONDITIONS ............................................................................................... 34 3.7.1 Birth registration and National ID cards ........................................................................................ 34 3.7.2 Piloting of Co‐responsibilities within the CT‐OVC ........................................................................ 35 3.7.3 Complementary interventions to support poverty reduction ................................................. 36 3.8 GENDER ................................................................................................................................................................... 37 4 INSTITUTIONAL ARRANGEMENTS ................................................................................................ 38 4.1 DESCRIPTION OF EXISTING INSTITUTIONAL ARCHITECTURE ...................................................................... 38 4.1.1 Overarching National Level Management ...................................................................................... 38 4.1.2 Institutional Arrangements of MGCSD‐Managed Programmes ............................................ 39 4.1.3 HSNP ................................................................................................................................................................. 40 4.2 ANTICIPATED CHANGES TO THE INSTITUTIONAL ARRANGEMENTS ........................................................... 41 4.2.1 Changes Resulting from the National Social Protection Policy and the Launch of the NSNP 41 4.2.2 Ministerial Restructuring ........................................................................................................................ 43 4.2.3 Devolution ...................................................................................................................................................... 43 4.3 INCENTIVES............................................................................................................................................................. 44 4.4 ASSESSMENT OF ADEQUACY ............................................................................................................................... 45 4.5 MANAGEMENT OF ERROR, FRAUD, AND CORRUPTION ................................................................................. 46 5 BUDGET PROCESS AND EXPENDITURE FRAMEWORK ........................................................... 49 5.1 BUDGET PROCESS AND STRUCTURE .................................................................................................................. 49 5.2 PROGRAMME FINANCIAL SUSTAINABILITY ...................................................................................................... 51 5.3 FUNDING PREDICTABILITY .................................................................................................................................. 52 5.4 ADHERENCE OF THE BUDGETED PROGRAMME EXPENDITURE AND EXECUTION TO GOVERNMENT PRIORITIES ......................................................................................................................................................................... 52 5.5 EFFICIENCY OF PROGRAMME EXPENDITURES ................................................................................................. 53 6 RESULTS FRAMEWORK .................................................................................................................... 54 6.1 OVERVIEW OF RESULTS FRAMEWORK ............................................................................................................. 54 6.1.1 Existing Results Frameworks ................................................................................................................ 54 6.1.2 NSNP Results Framework ....................................................................................................................... 55 6.2 MONITORING AND EVALUATION SYSTEM ........................................................................................................ 56 6.2.1 Existing Programme Arrangements .................................................................................................. 56 6.2.2 Proposed Arrangements for NSNP ...................................................................................................... 60 6.3 CAPACITY REQUIREMENTS .................................................................................................................................. 61 7 ECONOMIC JUSTIFICATION ............................................................................................................. 62 7.1 RATIONALE FOR PUBLIC PROVISION AND FINANCING .................................................................................. 62 7.2 THE PROGRAMME’S ECONOMIC IMPACT .......................................................................................................... 62 7.3 WORLD BANK ADDED VALUE............................................................................................................................. 63 7.4 RESULTS OF ECONOMIC EVALUATION .............................................................................................................. 63 8 INPUTS TO THE PROGRAMME ACTION PLAN ............................................................................ 64 8.1 TO CREATE MORE ROBUST SYSTEMS FOR TARGETING, REGISTERING BENEFICIARIES, MAKING PAYMENTS, AND MONITORING PERFORMANCE TO STRENGTHEN THE OVERALL GOVERNANCE OF THESE PROGRAMMES ..................................................................................................................................................................... 64 8.2 TO GRADUALLY HARMONIZE THE FIVE CASH TRANSFER PROGRAMMES TO INCREASE THE COHERENCE OF THE SECTOR .................................................................................................................................................................. 66 8.3 TO EXPAND COVERAGE OF THE FIVE PROGRAMMES IN A COORDINATED MANNER THAT WILL PROGRESSIVELY REALIZE THE RIGHT TO SAFETY NET SUPPORT.............................................................................. 67 9 TECHNICAL RISK RATING ................................................................................................................ 67 ................................ 68 10 INPUTS TO THE PROGRAMME IMPLEMENTATION SUPPORT PLAN REFERENCES ............................................................................................................................................... 70 ANNEX 1: RESULTS FRAMEWORK FOR NATIONAL SAFETY NET PROGRAMME ................. 74 Acronyms ASAL Arid and Semi-arid Lands BWC Beneficiary Welfare Committee CSO Civil Society Organization CTR Cash to Transfer Ratio CT Cash Transfer DCO District Children’s Officer DFID Department for International Development DLI Disbursement-linked Indicator DOGSD Department of Gender and Social Development DOSC District OVC Steering Committee DR Dependency Ratio EFC Error, Fraud, and Corruption FBO Faith-based Organization FEWS Famine Early Warning System GOK Government of Kenya HH Household HSNP Hunger Safety Net Programme ICT Information and Communication Technology IPRS Integrated Population Registration System KENAO Kenya National Audit Office KIHBS Kenya Integrated Household Survey KNBS Kenya National Bureau of Statistics Ksh Kenyan Shillings LOC Location OVC Committee MGCSD Ministry of Gender, Children, and Social Development MIS Management Information System MSDNKAL Ministry of State for Development of Northern Kenya and Other Arid Lands MTEF Medium-term Expenditure Framework MTEFF Medium-term Expenditure and Financing Framework NDDCF National Drought and Disaster Contingency Fund NDMA National Drought Management Agency NSNP National Safety Net Programme NSPC National Social Protection Council NSPP National Social Protection Policy OPCT Older Persons Cash Transfer OPM Oxford Policy Management OVC Orphans and Vulnerable Children PAEC Per Adult Equivalent Consumption PCK Postal Corporation Kenya PMT Proxy Means Testing PSNP Productive Safety Net Programme PSP Payment Service Provider PWSD-CT Persons with Severe Disability Cash Transfer RFM Risk Financing Mechanism SP Social Pension UFS-CT Urban Food Subsidy Cash Transfer UNICEF United Nations Children’s Fund WFP World Food Programme 1 INTRODUCTION 1.1 Context 1. Kenya has experienced a decade of relatively strong economic growth. Between 2000 and 2009, economic growth in Kenya averaged 3.7 percent. This translated into a modest rise in incomes for Kenyans – the first since the 1970s. However, growth declined sharply in 2008 and 2009 as a result of the violence following the December 2007 presidential elections, of the global food, fuel, and financial crisis, and of the drought that occurred after the fourth consecutive year with no rain during the usual rainy season. Economic growth rebounded strongly in 2010 and 2011, reaching 5.8 percent and 4.4 percent respectively. This strong growth is partly the result of the fact that the Kenyan economy, unlike many other African economies, is not principally driven by agriculture. Roughly 55 percent of its gross domestic product (GDP) comes from services, in which the private sector plays an especially dynamic and entrepreneurial role. Kenya has consolidated its position as a regional economic hub for many services, including transport, telecommunications, financial services, and information and communication technology. 2. Despite these advances, high rates of poverty persist in Kenya and unemployment remains a major concern, especially among urban youths. In 2005/06, poverty incidence was 46.6 percent, having declined from 52.2 percent in 2000. Poverty rates are markedly higher in rural areas (49.7 percent) than in urban areas (34.4 percent), although this lower overall urban poverty rate masks the great deprivation that exists in informal urban settlements. Differences can also be seen among administrative divisions, with 74 percent of the population living below the poverty line in the North Eastern Province compared with 22 percent in Nairobi Province. Finally, poverty rates also tend to be higher among vulnerable groups such as children (53.5 percent), particularly orphans and vulnerable children (54.1 percent), older people (53.2 percent), and people with disabilities (57.4 percent).1 3. This poverty does not just affect income and consumption. Children living in households in the poorer quintiles are less likely to attend school, tend to achieve lower levels of grade advancement, are less likely to benefit from vaccination programmes, and have poorer health outcomes than their better-off counterparts. The net enrolment rate2 for secondary schooling is only 6 percent for those in the bottom quintile compared with 52 percent for those in the top quintile. Only 54.8 percent of children from the poorest quintile have received a measles vaccine compared with 88 percent of the wealthiest, and the prevalence of stunting in the poorest quintile is 44.2 percent compared to 24.5 percent in the wealthiest quintile.3 Poverty in urban areas is characterized by poor hygiene and sanitation, high levels of morbidity, high child mortality, insecure housing tenure, and limited economic opportunities.4 4. Poverty in Kenya is intertwined with the continued vulnerability of the population to shocks. Between 2004 and 2012, the number of people in need of emergency assistance never fell below 900,000. The 2011 drought in the Horn of Africa again demonstrated the vulnerability of livelihoods in Kenya, leaving more than 3.8 million people in need of emergency assistance. Households have long reported experiencing shocks, the most significant of which were drought and death in the family, with poor households experiencing the negative effects of such shocks more often than others.5 Poor households were also more likely to resort to coping mechanisms that have long-term negative implications for their well-being, such as withdrawing their children from school or selling off their productive assets. This persistent poverty and vulnerability highlights the fact that social 1 Ministry of State for Planning, National Development, and Vision 2030 (2012) 2 The net enrolment rate is the number of pupils within the correct age group for education (in this case secondary education) enrolled in education expressed as a percentage of the total population in that age group. 3 World Bank, (2008) and KNBS and ICF Macro (2010) 4 OPM (2012b) 5 For example, extremely poor households were 78 percent more likely to report experiencing a negative effect of a shock than their wealthier counterparts (Ndirangu, 2010). 1 protection has an important role to play in the effort to reduce poverty and vulnerability and promote human capital development in Kenya. 5. Kenya has a long history of investing in social protection and has a number of well- established social insurance schemes and safety net programmes, but their coverage has tended to be low and their effectiveness limited. Most safety net support for poor and vulnerable populations is offered in the form of humanitarian relief (often as food aid), which has been mobilized by the government and the international community in response to crises such as drought and floods. In many parts of the country, most notably in arid and semi-arid lands (ASALs), this type of response has become common, with emergency food relief being provided year after year to chronically poor and food-insecure populations. Despite this long history of investment, the Government of Kenya has only recently (June 2011) developed a National Social Protection Policy. This policy builds on the Constitution of Kenya (2010) which includes in its Bill of Rights the “right for every person… to social security and binds the State to provide appropriate social security to persons who are unable to support themselves and their dependents.”6 6. Expenditure on safety nets in Kenya has risen rapidly in recent years. Between 2005 and 2010, total spending on safety nets increased from Ksh 11.9 billion to Ksh 20.5 billion. In 2010 this was equivalent to 0.80 percent of Kenya’s GDP. Throughout this period, emergency food aid continued to dominate safety net spending, accounting for 53.2 percent of the total. However, there has also been a growing trend towards cash transfers to the extent that the majority of government financing to safety nets has been spent on cash transfers in recent years. 7. As a result, the coverage of cash transfer programmes has grown significantly but remains low in comparison with the population in need. The Government of Kenya’s five main cash transfer programmes ‒ the Older Persons Cash Transfer (OPCT), the Cash Transfers to Orphans and Vulnerable Children (CT-OVC), the Hunger Safety Net Programme (HSNP), the Urban Food Subsidy Cash Transfer (UFS-CT), and the Persons with Severe Disability Cash Transfer (PWSD-CT) ‒ have collectively increased their coverage more than tenfold since 2005 and currently provide regular support to 1.6 million people in 279,843 households or 4 percent of the population. Despite these gains, cash transfers cover only a fraction of the absolute poor population (an estimated 46.6 percent of the total population). 1.2 Programme Description 1.2.1 GOVERNMENT PROGRAMME 8. Amongst the priorities outlined in the National Social Protection Policy (NSPP), is the government’s ambition to enhance social assistance by developing the necessary institutions and strengthening operational systems while expanding the coverage of such programmes.7 To help to realize this policy objective, the government proposes the establishment of a National Safety Net Programme (NSNP) by creating a framework around which the five principle cash transfer programmes will increasingly be coordinated and harmonized.8 These five programmes are described briefly below: 9. The Cash Transfers to Orphans and Vulnerable Children was launched in 2005 by the Government of Kenya in response to the growing levels of poverty and vulnerability caused by the increasing numbers of orphaned and vulnerable children (OVCs). OVCs are defined, for the purposes of the programme, as children who have lost one or both parents, are chronically ill or who have a caregiver who is chronically ill; and/or live in child headed households due to orphanhood. The 6 MGCSD (2011) 7 MGCSD (2011) 8 The government’s definition of social assistance is broader than the five programmes which make up the National Safety Net Programme. It includes school feeding, emergency assistance, cash and food based public works programmes and grants to vulnerable groups. 2 programme aims to improve the welfare of children living in poor OVC households while supporting poverty reduction and income distribution among the poorest in society. The programme’s main activity is the provision of two-monthly cash payments to poor households caring for OVCs. A small pilot is also in operation testing the appropriateness and possible impact of combining cash transfers with conditions, such as the condition that benefiting households take their children to school or to health centres. 10. The Hunger Safety Net Programme is an unconditional cash transfer programme that aims to reduce poverty in northern Kenya, by delivering regular cash transfers extremely poor households in four districts of Northern Kenya. The programme was designed in 2007 in response to the growing levels of chronic food insecurity found in Kenya, and in particular the arid, and to a lesser extent semi-arid, lands. The first phase of the programme was a pilot phase seeking to test which of three targeting approaches were most effective at reaching poor households, identify an effective payment system for delivering transfers reliably and safely and demonstrate the effectiveness and impact of cash transfers in fighting hunger and poverty. The programme is currently incorporating lessons learned in this pilot phase in the design of a second phase. As with all of the five cash transfer programmes it makes payments every two months. 11. The Older Persons Cash Transfer, launched in 2006, was designed in recognition of the fact that older persons constitute a sizeable poor population and are often poorer than other age groups. Furthermore, those entering old age in poverty are likely to remain poor, as there are fewer opportunities for them to engage in activities to alleviate their poverty. The programme aims to strengthen the capacities of older persons and improve their livelihood. It targets extremely poor households that include a member aged 65 or older who is not already receiving a pension. 12. The People with Severe Disability Cash Transfer aims to provide support to people with severe disabilities who are unable to look after themselves and require constant attention of caregivers. Beneficiaries include those who need permanent care, which includes but is not limited to feeding and hygienic care, and require protection from themselves, others, and the environment. The objective of the programme is to provide immediate relief to persons with severe disabilities from extreme poverty while enhancing their basic rights. 13. The Urban Food Subsidy Cash Transfer was conceived in 2008 as a response to extremely high levels of food price inflation. The objective of the programme is to improve the livelihood security of the most vulnerable residents of urban informal settlements in response to cumulative shocks and stress. The programme has a focus of increased access to food by vulnerable households in the short-term and a longer-term commitment to support such households to engage in longer-term food and income security initiatives. Although conceived in 2008, the Government programme was not launched until 2011. In the interim a number of initiatives were implemented by WFP, Oxfam and Concern. 14. The main objective of the NSNP, as laid out in the programme results framework (see Annex 1), is to improve the welfare of and increase resilience among specific vulnerable groups in order to reduce poverty and vulnerability in Kenya. Their vulnerability may be a result of where they live (the arid lands of Northern Kenya or in informal settlements of major urban centres) or their circumstances (households caring for orphans and vulnerable children, people with severe disabilities, and the elderly). This objective is to be achieved by delivering an efficient and effective safety net for poor and vulnerable households in the form of the five cash transfer programmes listed above. The principal activities undertaken by these five programmes are outlined in Table 1 below according to programme output: 3 Table 1: NSNP Outputs and Activities Outputs Activities 1 The most vulnerable and Targeting (both geographical and household), recertification, and poorest households are change management9. This includes gradually expanding enrolled in the NSNP. programme coverage 2 NSNP beneficiaries receive Implementation of payment mechanisms that deliver timely appropriate, reliable, and transfers to beneficiaries while maintaining proper controls and accessible payments. regular reviews of the benefit amounts to ensure that they are adequate 3 Citizens are able to appeal Management of functioning grievance mechanisms and greater and complain to improve awareness of programme objectives and modalities by both programme performance. beneficiaries and non-beneficiaries 4 Programmes are responsive Putting in place the capacity to scale up transfers in response to to disaster. shocks and ensuring appropriate links to existing early warning and disaster response capacity 5 A monitoring and learning Development of a management information system for each system is functioning. programme and an overall single registry that collates information from all of the programmes as well as putting into operation other key monitoring and evaluation tools 6 The cash transfer sector is Development and implementation of a plan to consolidate cash harmonized and is transfer programmes and the inclusion of all expenditures under government-owned. these five programmes in the budget 15. Initially most of these activities will be carried out on a programme-by-programme basis, but as the NSNP proceeds, the government will exploit any opportunities to share functions across programmes as outlined in output 6. 16. The programmes managed by the Ministry of Gender, Children and Social Development (CT- OVC, OPCT, UFS-CT and PWSD-CT) follow the general government administrative structure with national-level management and oversight, a small amount of monitoring and supervision by province- level staff, district-level coordination of implementation and village, and community-level involvement. At the national level, responsibility for implementing these programmes lies with either the Division of Social Welfare of the Department of Gender and Social Development (for the OPCT, UFS-CT and the PWSD-CT) or the Department of Children’s Services (for CT-OVC). 17. The HSNP is the responsibility of the HSNP Secretariat within the Ministry of State for Development of Northern Kenya and other Arid Lands (MSDNKAL). Although the secretariat is based within MSDNKAL, the staff are short-term contractors financed by the development partner. Furthermore key areas of the HSNP’s implementation, such as targeting, management of appeals and grievances, payments and monitoring and evaluation, are sub-contracted to NGOs and private sector partners. 18. The government’s expansion10 of cash transfers will increase the number of poor benefiting from the current 1.6 million to 2.89 million in 2016/17. This increased coverage will require additional financial resources. The current allocation to these programmes is Ksh 8.04 billion. To achieve the scale-up targets proposed, the five programmes will require a total of Ksh. 10.8 billion in FY 2013/14 and Ksh. 17.4 billion by 2016/17. The government also plans to progressively take on financing to these programmes. As the NSNP begins in 2013/14, GOK financing will represent 9 Change management refers to the management of information concerning changes to households including births, deaths, changes of address, update of national identification numbers etc. 10 The figures referred to in this paragraph are drawn from the current version of the programme Medium Term Expenditure and Financing Framework (Version 9 April 2nd). This framework will continue to be reviewed and revised over the coming months. 4 around 40 percent of the programme financing. The GOK will progressively increase resourcing to the programmes so that by 2016/17, its allocation will have increased substantially from Ksh. 4.3 billion in 2013/14 to 13.5 billion in 2016/17. With this plan, the government will provide 77 percent of financing to the cash transfer programmes. 19. As discussed in detail in section 5.2 below, the NSNP will cost less than 0.5 percent of GDP (with government spending on the programme less than 0.3 percent of GDP) and will constitute between 1 and 1.4 percent of government expenditure. If the government’s growth projections are fulfilled and revenue increases as high as anticipated, this level of spending will be affordable.   1.2.2 WORLD BANK PROGRAMME 20. This proposed Program for Results (P4R) operation aims to support the government’s ambitious objective of establishing and progressively expanding the National Safety Net Programme. The boundaries of the World Bank’s programme are defined as all activities under the five cash transfer programmes as described above in Table 1 but will be focused on supporting the government to achieve the following three objectives: 21. To create more robust systems for targeting, beneficiary registration, payments, and monitoring to strengthen the governance of these programmes. World Bank support will assist the government to adopt common standards for all cash transfer programmes that will strengthen programme governance. This will entail the adoption by all programmes of the emerging best practices in the sector, which include: (i) strengthening programme targeting to ensure that only those households that are eligible for the programme are selected; (ii) establishing a Single Registry to strengthen internal controls on enrolment, thereby providing additional confirmation that households registered in the programmes are eligible for support; (iii) adopting internal controls on the payrolls produced through the electronic programme MISs to better ensure that the correct amount is paid to beneficiaries; (iv) contracting payment service providers that make payments electronically and use two-factor authentication to ensure that payments reach the intended beneficiary in an efficient and effective manner; and, (v) establishing well-functioning grievance and appeals systems to strengthen demand-side accountability in the programmes. 22. To increasingly harmonize the five cash transfer programmes to improve the coherence of the sector. The current approach to implementing cash transfers creates overlaps, duplication of effort and inefficiencies. To address these limitations the MGCSD, with support from the World Bank, will formulate a strategy for consolidating its four cash transfer programmes. This plan will include: (i) an agreed proposal for merging (or harmonizing) some of the main functions of the cash transfer programmes; (ii) the institutional arrangements at all levels; (iii) the staffing requirements; (iv) the physical resources requirements; and (v) any policy, regulatory or institutional changes required. Furthermore, the HSNP programme is dependent on development partner financing which is “off- budget”. Safety nets require multi-annual commitments and, in order to be sustainable and allow oversight, need government financing and should be reflected in the government’s budget document. The P4R will support the government to reflect all financing to the NSNP in its budget and begin to allocate resources to the HSNP. This objective also focuses on strengthening regular, comprehensive programme reporting and links between the programme’s Single Registry and the government’s Open Data Initiative. Furthermore, coordination between the cash transfer programmes and emergency response will be enhanced through efforts to build mechanisms to scale up cash transfers in response to shocks. 23. To expand the coverage of the five programmes in a coordinated manner that will progressively realize the right to safety net support. This objective is concerned with both geographic expansion of the five programmes and increasing household coverage within programme areas. Presently, increases in programme size are driven by available budget and little consideration for the coverage (and expansion plans) of the other programmes. The World Bank and government have agreed to address this by promoting the adoption of a consolidated scale-up plan for the NSNP that 5 will be based on transparent criteria on poverty and vulnerability, based on the most recent census (2009). 24. The key results for the World Bank operation are:  The net change in the monthly consumption expenditure among beneficiary households, as a reflection of the positive impact of the program on beneficiaries’ welfare.  Number of NSNP beneficiaries, showing the increasing coverage of the NSNP.  Percent of NSNP beneficiaries who conform to program targeting criteria, reflecting improvements in the implementation of the targeting method  Percent of payments disbursed to payment service providers on time, reflecting the delivery of timely payments to beneficiaries. 2 STRATEGIC RELEVANCE 25. The strategic relevance of this intervention is assessed below according to three key criteria:  The appropriateness of the intervention to the country’s needs  Its alignment with the Government of Kenya’s priorities  The opportunities presented by investing in the sector. 2.1 Appropriateness of the Intervention to Needs 26. As stated above, high rates of poverty persist in Kenya despite strong economic growth. As a consequence, inequality is growing and there is an increasing gap between the consumption levels of those in the richest and poorest deciles (the national consumption decile ratio11 rose from 13 to 19 between 1997 and 2005/6). Comparing average consumption in different areas further highlights this inequality. The average expenditure of those in the bottom rural decile is only 8 percent of the expenditure of those in the top rural decile, and in urban areas the ratio is 1:20.12 27. The five cash transfer programmes that constitute the NSNP focus on groups that the government has identified as being particularly vulnerable 13 . These groups include orphans and vulnerable children, those living in drought-affected arid lands, the elderly, people with severe disabilities, and those living in informal settlements in major towns. As Table 2 below shows, these groups experience higher levels of poverty than the rest of the population. 28. Cash transfers have a proven track record internationally, and in Kenya, in reducing poverty and inequality. Such programmes have enabled targeted households to spend more on household necessities such as food, fuel, and housing and to invest more in their children’s health, nutrition, and education. These effects are stronger in areas and amongst groups which experience poverty more severely. These effects are explored further in section 7. 11 The consumption decile ratio expresses the income of the richest decile as a multiple of those in the poorest decile. 12 World Bank (2008) 13 These are not the only groups listed as vulnerable in the Kenyan Constitution (National Council for Law Reporting (2010)). Other defined vulnerable groups include women, youth and members of marginalized communities. 6 Table 2: Population and Poverty Rates of Vulnerable Groups in Kenya 2.2 Alignment with the Government of Kenya’s Priorities 29. The new Constitution aims to move Kenya towards a more equitable and inclusive future. Specifically, the Constitution states, “Every person has a right to social security… [and] the State shall provide appropriate social security to persons who are unable to support themselves and their dependants.”14 The National Social Protection Policy (NSPP), which was recently approved by the Cabinet, outlines a vision for progressively realizing this right to social protection. 30. Vision 2030 specifies the need to increase financing for safety nets as a way to reduce poverty and promote economic growth. The objectives laid out in Vision 2030, in the NSPP, and the Constitution will inform the development of the government’s next mid-term plan. 31. As already stated above, Kenya has a long history of investing in social protection. It has a number of well-established social insurance schemes and safety net programmes, and the coverage of its cash transfer programmes has increased significantly in recent years. By the end of 2010, cash transfers constituted 25 percent of total safety net spending in Kenya, having increased rapidly from a very low base in 2005, and now account for the majority of government spending on safety nets.15 14  The right to social security includes both social assistance and social insurance and is thus synonymous with social protection.  15 Ministry of State for Planning, National Development, and Vision 2030 (2012) 7 Figure 1: Government Financing for Safety Net Programmes, 2005-2010 Notes: Social cash transfers refer to the five main cash transfer programmes. Relief and recovery refers to programmes such as emergency food aid and supplemental feeding. Education refers to regular school feeding programmes. Health refers to feeding programmes and vouchers for primary health services. Agriculture refers to programmes that provide inputs and small grants to farmers’ groups. 32. The National Social Protection Policy16 recognizes the need to increase the scale and improve the quality and coordination of social assistance and lays out the government’s commitments in this area. On social assistance, it commits to strengthening and scaling up existing social assistance programmes, establishing an institutional framework for coordinating social assistance to ensure consistent and adequate levels of support, and establishing the funds and institutions needed to provide assistance. It is out of these commitments that the National Safety Net Programme (NSNP) was developed.   2.3 The Opportunities Presented by Investing in the Sector 33. Kenya has made rapid progress in recent years in increasing investments in and coverage of cash transfer programmes. However, the sector remains fairly new, and three of the five cash transfer programmes operate on a very small scale. Despite recent progress, there continues to be significant fragmentation in the sector. Cash transfer programmes are implemented in three different departments and two different ministries. As a result, there is limited coordination between programmes and a duplication of effort, which constitutes a missed opportunity for optimizing the use of limited implementation capacity and for ensuring a coherent approach to capacity-building across the sector. Furthermore, the coverage by these programmes of vulnerable groups and the absolute poor remains inadequate. Without further investment in the sector, there is a risk that coverage will remain low, that the sector will stay fragmented, and that several of the existing programmes will never function on a large enough scale to achieve measurable results. The NSNP was conceived as a response to these challenges and to achieve the objectives of the National Social Protection Policy to strengthen operational systems while expanding the coverage of cash transfers. 34. The five programmes that make up the NSNP share similar objectives and implementation modalities, including the need for targeting, payment systems, and monitoring and evaluation frameworks. This makes a more coordinated approach both feasible and desirable. Increasing coordination among these programmes will make it possible to ensure that effective safety net support is provided to the vulnerable population at the sector level, while adopting common implementation systems and procedures will increase the efficiency of programme delivery. 16 MGCSD (2011) 8 35. The NSNP will coordinate the expansion of the five programmes until together they achieve comprehensive and equitable coverage of the vulnerable population. The NSNP will require the managers of all five programmes to produce multi-year targets for progressively increasing their beneficiary numbers. Each programme will prioritize geographic areas based on agreed poverty and vulnerability criteria. However, the coverage of cash transfer programmes will need to be expanded in a way that will ensure the long-term sustainability of the NSNP. This will require government financing to progressively increase in relation to development partner financing and implementation capacity to be strengthened. Therefore, the pace at which the cash transfer programmes are able to expand their coverage will depend, in large part, on the pace at which the government is able to increase its safety net budget and increase implementation capacity. 36. In support of the NSNP, the P4R will entail the adoption by each of the five programmes of the emerging best practices in the sector. These include (i) electronic management information systems (MISs) that can process consolidated programme payrolls; (ii) payroll systems that use agreed standards for internal payroll controls;17 (iii) payment service providers that use two-factor authentication; and (iv) well-functioning complaints and appeals systems. Furthermore, the programme will support the development of a single registry, a common database on which all information on the beneficiaries of all five cash transfer programmes will be consolidated. This registry will have several advantages. It will: constitute a consolidated source of information on cash transfer programmes for policymakers; enable the cross-checking of beneficiary lists to eliminate any “ghost” beneficiaries; reduce the possibility of individuals benefitting from multiple cash transfer programmes; enable the rapid identification of potential beneficiaries in need of emergency assistance; and be a foundation for creating other common delivery systems, including payment systems, which will reduce the cost of delivering benefits. Proposed increases in coverage by the programmes which make up the NSNP are being presented in the 2013 Government of Kenya Medium Term Plan.18 37. Overall the NSNP, and the P4R which supports it, can be seen as an appropriate response to the high rates of poverty experienced in Kenya and a timely intervention given the government’s commitment to social protection in general and cash transfers in particular. It also provides a significant opportunity to strengthen the quality and enhance the coherence of interventions within the sector. This investment is critical in helping the government to change the current situation characterized by a series of fragmented programmes with varying levels of capacity into the beginnings of a sustainable national system of social assistance provision. 3 TECHNICAL SOUNDNESS 38. A technically sound programme is one that is designed and implemented to efficiently yield the required results and achieve its objectives. This can be assessed on the basis of programme experience to date and in relation to international experience and best practices. The key parameters for assessing whether a safety net programme is technically sound are as follows:  The programme’s scale and scope is commensurate with programme objectives and is affordable  Targeting procedures ensure that that those who are in need and are entitled to support from the programme, receive this support  There are adequate information systems for keeping track of programme beneficiaries and for supporting programme operations 17 The internal payroll controls will consist of: (i) Pre-payment payroll checks by each programme MIS that analyses the list of beneficiaries enrolled in the programme that are to be paid in that payment cycle by household ID numbers, caregiver/beneficiary names and locations to identify any duplicate entries and cross- references the entitlement of the household under the programme with the amount to be paid; and (ii) Ex-post payroll checks to analyse the amounts actually paid to beneficiaries to ensure that payments were complete and accurate. 18 MGCSD (2013) 9  Beneficiaries are receiving adequate, reliable, and accessible payments  Beneficiaries and citizens are able to raise their concerns about programme delivery and can influence the service that they receive and future improvements in the programme  The programme is appropriate for the context in which it is implemented and is designed to be scaled up in response to shocks in ways that support the achievement of the programme’s objectives  Adequate provision has been made to ensure that programme beneficiaries receive any other services that they require in order for programme objectives to be achieved  The programme has taken account of gender differences to ensure that the impact of the programme is maximized. 39. The sections below outline how the programmes which make-up the NSNP address these parameters, assesse the sufficiency of these actions and identify key areas where improvements are desirable. Critical improvements are further summarised in section 8. 3.1 Scale, Scope, and Target Group 40. The five cash transfer programmes that constitute the NSNP focus on groups identified by the government as being particularly vulnerable. These groups include orphans and vulnerable children, those living in drought-affected arid lands, the elderly, people with severe disabilities, and those living in informal settlements in major towns. As stated above, these groups experience higher levels of poverty than the general population and also have less access to social services and are more likely to be socially marginalized. Table 3 below summarizes poverty levels, numbers of absolute poor households and planned coverage levels by differing target groups. 41. By the end of its first phase (2013-2017), the NSNP aims to provide support to 534,054 households, reaching a total population of approximately 2,970,211 individuals. This population will constitute approximately 15 percent of the absolute poor and 22 percent of the absolute poor within the targeted vulnerable groups (assuming perfect targeting and ignoring the likely double-counting) and will be a significant increase over current coverage levels. However, the high number of eligible households remaining uncovered indicates that further work is needed for the safety net to be effective. This highlights the need for the government to further refine targeting criteria to limit the eligible population and/or to further expand the programme in the long-run. 42. The NSNP identifies vulnerable households using a combination of categorical and geographical targeting. It distinguishes specific categories of households that are widely recognized as being vulnerable and in need of social assistance. The geographically defined groups are those who live in areas with markedly higher poverty levels and that have historically been targeted to receive assistance in response to shocks (whether droughts or food price shocks). The categorically defined groups include those experiencing higher than average consumption poverty and affected by other deprivations or increased needs. Both of these groups represent those who have clear and recognized vulnerabilities and are large enough to allow for some narrowing down of the target population using poverty criteria. Other possible categories of vulnerable groups are unlikely to significantly narrow down the target population prior to poverty targeting (for example, households with children, which constitute 73 percent of the total population). However, it may be possible to explore alternative sub- categories such as households with children aged 0-2. 10 Table 3: Numbers of Absolute Poor Households Within Programme Target Groups 3.2 Targeting, Change Management, and Recertification 3.2.1 TARGETING 43. Targeting resources to those households that need them most can maximize the impact of a programme within a specified budget. Therefore, the targeting system is a key component of any social protection programme. It ensures that those who need and are entitled to the benefits provided by a programme receive this support. There are two key aspects of any targeting system:  Defining what group of the population should benefit from an intervention (this group might be defined on the basis of poverty or some other form of vulnerability such as old age or disability or a combination of both).  Implementing a system that successfully identifies and enrols those who are a member of the above group. 44. The principles that should govern a strong targeting system are:  It maximizes coverage of the target group  It minimizes leakage to people who are not members of the target group  It is cost-effective (the costs of improving targeting quality do not outweigh the efficiency gains)  It is transparent in ways that enhance credibility and reduce fraud.19 19 Derived from Castañeda et al (2005) 11 45. The most common targeting approaches are geographical targeting, targeting of specific categories of the population (children or the elderly), and poverty-based targeting. The five cash transfer programmes use a combination of these approaches (including a number of different poverty targeting methodologies) to identify eligible beneficiaries. 3.2.1.1 Geographic Targeting and Resource Allocation Formula 46. Geographic targeting is commonly used when there are significant variations in poverty levels in different areas of the country. It may be used to exclusively target certain areas to the exclusion of other, less vulnerable areas, or it may be used to allocate higher levels of resources to needier areas and lower levels to less needy areas. 47. There are significant variations in the poverty levels of counties, districts, and even locations within districts as illustrated in Figure 2. Figure 2: Percentage of the Population in the Bottom Quintile of Expenditure in Selected Districts Source: KNBS (2007) 48. There are also some variations, though less dramatic, in the number of members of categorical groups in different parts of the country. For example, districts located in western Kenya seem to include a higher proportion of OVCs than districts elsewhere in the country, while the prevalence of reported disability appears to be higher in Nyanza Province (6.8 percent) than the national average (4.6 percent).20 49. Most programmes initially employed some form of geographical targeting (with the exception of the PWSD-CT, which supports approximately 70 beneficiaries in each constituency). The UFS-CT and the HSNP continue to operate in defined geographic areas (Mombasa city and four counties in Northern Kenya respectively), while both the CT-OVC and the OPCT were initially rolled out in those districts that had been identified as being the most vulnerable. The manuals of both the OPCT and the CT-OVC specify that these programmes must use geographic poverty-based targeting, with the CT-OVC having a clearly defined mechanism (see Box 1 below) for doing so. However, after this initial period, the OPCT stopped using geographic targeting on the basis of poverty but instead has responded to requests from Parliament to try to achieve standard levels of coverage in all constituencies. 20 NCAPD and KNBS (2008) 12 Box 1: A Comparison of the Geographic Targeting in the CT-OVC and OPCT CT-OVC OPCT The programme uses data from the 2005/06 Kenyan The OPCT was initially launched in 2007 in three Integrated Household Budget Survey (KIHBS) to rank the districts and quickly expanded to 44 districts by 69 districts according to the number of extremely poor 2009, supporting 750 beneficiaries in each district. households that contain OVCs. Once a district is selected, At this time, the programme covered 79 programme managers are expected to consider all constituencies, leaving 131 constituencies constituencies in the district as eligible for support. The unsupported. number of extremely poor households that contain OVCs In response to pressure from Parliament, is also used to determine the number of locations policymakers decided to expand to the remaining supported in each constituency, with a minimum of one 131 constituencies. Because of resource constraints, location covered per constituency. only 23 beneficiaries per constituency were The 1999 census data are then used to rank locations supported. within each constituency according to the number of Further resources were made available this financial OVCs. This ranking exercise determines which locations year, and the programme is currently expanding a the programme supports first. Constituencies with only a second time into these remaining 131 small number of poor OVC households will only work in constituencies. Districts are being asked to identify the location with the largest number of OVCs, whereas an additional 180 potential households per those with larger numbers of poor OVC households will constituency, of whom 150 will be enrolled in the be able to expand into further locations21. programme while the remaining 30 will be placed While the number of OVCs per location acts as a planning on a waiting list. figure determining the extent of data collections in each Although the allocation per constituency is equal, location, the actual number of programme beneficiaries district-level policymakers are encouraged to per location or district depends on the outcome of the ensure that within each constituency more resources proxy means test mentioned below. are allocated to poorer locations (as indicated by As coverage of the CT-OVC expanded, policymakers locally available data). decided that it should operate in every constituency in the country. The allocation of how many locations can be supported in each constituency continues to be based on poverty data, as does the number of households supported in each location. The programme is now operational in all counties and all constituencies of the country, but with the number of beneficiaries dependent on levels of poverty and vulnerability. 50. The differing approaches to geographic targeting, combined with the budgets available for the different programmes, have significant consequences for the spread and depth of the five programmes that make up the NSNP. Some programmes (the UFS-CT and the HSNP) focus on covering all locations within a narrow geographic area and include a significant number of beneficiaries per location. Meanwhile others, in particular the PWSD-CT and OPCT, are spread throughout the country but have only a few beneficiaries in each location. Table 4 below summarizes the different spread and depth of the programmes’ coverage. 21 This approach undermines the ability of the programme to treat locations in different constituencies in an equitable manner. It is likely that the poorest location in a better off constituency will contain less poor OVC households than an excluded location in a very poor constituency. 13 Table 4: Spread and Depth of Programme Coverage Geographic Spread* Total in CT-OVC HSNP OPCT PWSD- UFS-CT Country CT No of Counties 47 47 4 47 47 1 No of Districts 69 69 4 69 69 4 No of Constituencies 210 210 14 210 210 4 No of Locations 2,427 697 136 No of Beneficiaries** 148,769 69,000 39,388 14,700 10,168 Notes: * When the cash transfer programmes were launched, districts, as the key administrative unit, were the major unit for programme planning, with locations a secondary unit for detailed planning within districts. In 2005 there were 68 districts and approximately 1,500 locations in the country. In recent years there have been major alterations to the district boundaries that have resulted in a significant increase in the number of districts (there are currently 286 districts in the country). Also counties have now been accorded new responsibilities under the constitution, and constituencies have grown in importance due to the growing authority of Parliament. Both are likely to feature more strongly as planning units in the future. The number of counties is limited to 47, and the number of constituencies was 210 but the upcoming election will be fought in 290 constituencies. **According to the most recent payroll. 51. As programmes have expanded, there has been an increased chance of overlap of two or more programmes in the same geographic area. As the table above demonstrates, there is already significant overlap at a constituency level with a lesser but still increasing overlap at location and sub- location level. There has been little guidance provided, to date, on how local level implementers should manage potential overlaps other than statements that households should not benefit from more than one cash transfer programme. 52. There is a need to improve the evidence used to decide which locations are prioritized and how resources are allocated at the county, district, constituency, and location levels to ensure that programmes are responding appropriately to the distribution of vulnerability and poverty. This is necessary to ensure political support for the programmes and to ensure that the resources spent are proportional to the poverty and vulnerability of different geographical areas. Discussions on how to do this are already under way between the MGCSD and the World Bank and involve using consumption data from the 2005/06 Kenya Income and Household Budget Survey combined with information from the 2009 Census to rank localities based on the absolute poverty headcount and estimate the total number of poor people in each vulnerable group eligible for each programme. 3.2.1.2 Household Targeting and Registration 53. The cash transfer programmes currently use a variety of different combinations of categorical targeting, subjective poverty judgments, community targeting and proxy means testing to target their beneficiaries. 54. For example, the CT-OVC uses a four-stage targeting process after the application of the geographical targeting exercise described above:  Community representatives compile a list of households that meet both of the programme’s categorical criteria - they contain at least one orphan or vulnerable child and they are poor22.  Enumerators employed by the programme collect detailed poverty status related information on these potential beneficiary households.  The national programme secretariat then applies a proxy means test to the information collected on this list of potential beneficiaries to rank them in terms of need, from which it produces a priority list of eligible households.  This priority list is then confirmed and validated at community meetings. 22 The CT-OVC defines a child as an OVC if s/he has lost one or both parents, is chronically ill or who has a caregiver who is chronically ill; and/or lives in a child headed household due to orphanhood. 14 55. This targeting approach was adopted after it was found that the previous, poverty score card, method, while successful in identifying households that contained orphans and vulnerable children, had been less successful in prioritizing the poorest households that met this criterion.23 An upcoming baseline survey24 is expected to provide an assessment of how successful this new approach has been in prioritizing poor households. 56. The HSNP tested three alternative targeting mechanisms in order to assess which was most effective at identifying poor and vulnerable households. Each sub-location was randomly assigned one of the targeting mechanisms, and the success of each approach was then evaluated using a combination of quantitative and qualitative analysis.25 The three approaches that were piloted were:  Community-based targeting (CBT). Communities were tasked with identifying both appropriate targeting criteria and the neediest households up to a specified quota. This approach was put into action with the intensive support of partner NGOs.  A universal social pension for all individuals above a certain age. A 55 year cut-off was used for those who could prove their age using official documents (such as a national identification card or a birth certificate), whereas the age of 60 was used as the threshold for those who had to prove their age using a historical calendar implemented by a vetting committee26.  A dependency ratio (DR) approach. This method identifies and targets households who have a high number of dependants - defined as children aged 17 years old and under, the elderly aged 55 years and over, and the disabled or chronically ill (with either physical or mental impairments). These cut-offs varied from district to district. 57. A 2011 assessment of the effectiveness of these three methods concluded that the CBT approach was the most effective for identifying the poorest households. Beneficiary households in CBT areas were 50 percent more likely to be poor (falling below the 51 percent relative poverty line) than non-beneficiary households compared to 14 percent more likely in the social pension areas and 17 percent in the dependency ratio areas. The performance of the dependency ratio approach was undermined by implementation errors, as 30 percent of households receiving the benefit were found not to be eligible (while 23 percent of eligible households were not covered). The targeting accuracy of the social pension was very high, demonstrating the relative ease of applying the age-based criteria (96 percent of beneficiaries were eligible and 83 percent of eligible households were covered by the programme). However, it appears that the presence of an elderly person in a household is not strongly associated with poverty in the programme’s operational areas.27 This assessment also assessed the possible results of using a proxy means test to target HSNP resources and found that there was a possibility of such an approach outperforming any of the tested methodologies. 58. The HSNP is currently revising its targeting approach to use a combination of community targeting and a proxy means test (PMT). The members of the team designing the approach have drawn on the PMT methodology used by the CT-OVC. The results of both approaches will be compared while targeting teams are in the field, and any differences in how households are ranked or grouped will be resolved by checking the accuracy of the data collected and by discussion between targeting teams and community leaders. The field-testing of the new methodology is currently under way. 23 OPM (2010) 24 A new baseline for the CT-OVC has recently been commissioned by the World Bank, along with follow-up evaluation surveys in order to assess the quality of programme implementation and impact over the next two years. This baseline covers districts not previously covered by OPM evaluations which were focused on operational areas financed by DFID and UNICEF support. 25 OPM and IDS (2011) 26 The historical calendar lists locally known historical events such as severe droughts, conflicts or other well known incidents. These are then compared with what is known about a potential beneficiaries birth and childhood to come up with an estimation of their year of birth. 27 OPM and IDS, (2011) 15 59. The PWSD-CT and OPCT both used to rely heavily on community leaders to identify eligible households. Although the programme manuals specified detailed targeting approaches, including the collection of data on a variety of indicators, only a limited amount of data was collected, which meant that programmes mostly used the subjective judgments of local leaders to identify the most eligible households. A recent review of these two programmes, conducted in three counties, found that this kind of targeting was largely successful in identifying households that met the categorical criteria but was less successful in prioritizing these households by poverty status. Fewer than 1 percent of OPCT respondents and 7.6 percent of PWSD-CT households reported not having a member meeting eligibility requirements. The higher figure for the PWSD-CT is largely made up of households with a members living with a disability, but the respondents in these cases did not categorise the disability as severe. However, according to the poverty scoring approach28 used in the review, the likelihood that the OPCT and PWSD-CT beneficiaries interviewed would fall below the poverty line is only between 30 and 37 percent for Nairobi and Embu (the figure is better in the third review location with as much as 67.9 percent of OPCT beneficiaries likely to fall below the poverty line). 29 60. More recently the OPCT has revised its targeting approach to include the use of a poverty scorecard. Under the new system, which started in November 2012, local informants or committee members now collect detailed household information that is then analysed, according to a poverty score card, by the team at the Department of Gender and Social Development (DOGSD) in Nairobi to produce a preliminary beneficiary list. It is planned that a similar poverty score card approach be used in future for both the PWSD-CT and OPCT programmes as well as the UFS-CT programme described below. It is too soon to be possible to assess the results of this approach, however it is likely that lessons could be learned from the more rigorous indexing of poverty indicators employed by the proxy means tests of the CT-OVC and HSNP. 61. The UFS-CT used a large-scale data collection exercise to identify eligible households. Village elders guided enumerators to vulnerable households within the community, and the enumerators then calculated each household’s per adult equivalent consumption (PAEC) on the spot. Households whose PAEC was below KSH2000 were then asked to respond to a full-length questionnaire. Programme administrators then applied a proxy means test to this information to establish a list of target households that were the most poor and food-insecure. As enumerators were assessed and paid on the number of questionnaires that they had completed, there was an incentive for them to distort the data used for the PAEC calculation to achieve a result of below KSH2000. In addition there were reports of village elders and enumerators demanding kick-backs from those households who were included in the programme. Despite these challenges, it was found that the households selected for the programme were actually poorer than average and tended to include more vulnerable members (such as OVCs, the disabled, and ill people).30 62. Most programmes include some kind of community engagement in the targeting process, particularly the HSNP. The other four programmes attempt to hold barazas (community meetings) at the location or sub-location levels to validate the lists of targeted households. It seems that this is largely implemented in the CT-OVC,31 but assessments of the OPCT, PWSD, and UFS-CT found that beneficiaries are frequently unaware that these meetings are taking place.32 Even when the barazas do take place, they are not well attended, and participants tend to be reluctant to challenge the beneficiary lists.33 28 The review scored households using the Grameen Foundation’s Poverty out of Poverty Index (PPI). The PPI works as a type of poverty score card by estimating the likelihood that a beneficiary has an income below a given poverty line. 29 Kimetrica (2013) 30 Concern (2012) 31 Calder et al (2011) 32 Kimetrica (2013) and Concern (2012) 33 Calder et al (2011) and Amuyunzu-Nyamongo (2012) 16 63. Many of the challenges experienced within the PWSD-CT, the OPCT, and the UFS-CT were the result of limited investment in implementation of targeting. District officers and community leaders reported that there were insufficient funds and time to properly implement the targeting methods. Although these challenges were worse in the PWSD-CT, OPCT, and UFS-CT, budget constraints were also reported in the case of the CT-OVC, and the limited time to complete targeting was highlighted as a problem for both the CT-OVC and the HSNP.34 64. In conclusion, most assessments confirmed that targeted beneficiaries largely meet the programmes’ criteria, particularly in relation to categorical criteria such as old age or orphanhood but less so in terms of their poverty level, and most programme beneficiaries reported that the targeting process was fair (ranging from 78 percent in the OPCT to 96 percent in the HSNP).35 However, a lower proportion of non-beneficiaries perceived the targeting process as fair (around 50 percent in the OPCT and the HSNP).36 Also, there are concerns about the large numbers of eligible households that are excluded from the programmes (which is inevitable given the limited budget and relatively high levels of need) and about the inadequacy of the targeting mechanisms for identifying those most in need from among all of the households that meet the categorical eligibility criteria. There have also been isolated reports of targeting fraud (particularly in relation to the UFS-CT) or of enumerators or community leaders demanding payments from households in return for including them in the programme.37 65. Reviews are currently underway of the OPCT, PWSD-CT and UFS-CT programmes, and these are likely to lead to revisions of the targeting procedures used in these three programmes. Recommendations focus on (i) improving the targeting criteria by introducing a more rigorous approach to poverty targeting to the OPCT and the PWSD-CT and refining the approach used by UFS-CT; (ii) simplifying the targeting form so that it only includes the data that will be used; and (iii) improving the documentation and training on targeting procedures (including the provision of guidance materials in Swahili). There is also a need to strengthen the following aspects of the targeting systems in all five programmes:  Community awareness. A lack of community awareness is frequently cited as one of the weaknesses in targeting. Local people are frequently not certain about whether or not they are eligible for the programme and about whether those making targeting decisions are doing so correctly. It is important to increase community awareness of programmes and of their intended target populations in the future to increase effective coverage and to hold service providers to account.  Targeting duration and oversight. Hurried implementation of targeting, particularly when combined with limited oversight, has caused problems in a number of programmes. As a result of rushed implementation, some households have been missed out while incorrect information has been recorded about others, all of which has led to mistakes in the selection of beneficiary households. A lack of sufficient support for and supervision of local implementers has also caused mistakes to be made with no opportunity to correct them. Reviews of the UFS-CT, PWSD-CT and OPCT point to the need to revisit a sample of screened households to verify the information collected. 38 There is a need to ensure that in future a sufficient amount of time is allowed for the proper implementation of the targeting process as well as an appropriate level of staffing to support and oversee local implementers.  Community validation. A number of evaluations have highlighted weaknesses in the use of community barazas for validating beneficiary lists. In some programmes, the barazas either 34 Calder et al (2011), OPM (2012a), and Concern (2012) 35 OPM and IDS (2011) and ADR (2011) 36 ibid 37 Concern (2012) noted reports by respondents of village elders and enumerators informing those they were registering that if they received the cash transfer then they should pay back KSH 2,000-3,000 of it to the village elders/enumerators. 38 Concern (2012) and Kimetrica (2013) 17 were not held or were held at the district level, which limited community participation. In other programmes, the barazas were held but acted more as rubber stamping exercises than as forums in which people actively discussed the eligibility of potential beneficiaries for the programme. Therefore, there is a need to strengthen the community baraza process in order that it can actually act as a verification exercise and is genuinely able to change targeting decisions.  A more comprehensive initial screening of target populations. In the categorically targeted cash transfer programmes, the identification of the initial population of target beneficiaries should be both exhaustive and well documented. This will make it less likely that households will be accidentally excluded and will make it easier to audit targeting processes to check that procedures have been properly followed. 66. Furthermore, as all programmes expand their coverage, there will be a growing number of counties and locations within which more than one programme is functioning. This is a key step in Kenya towards realizing the government’s ambition to provide safety nets for multiple vulnerable groups. It also makes it possible to use the same targeting system for multiple programmes. Apart from being a more efficient use of time and resources, this is also likely to reduce gaps in coverage and eliminate the risk of the same household becoming a beneficiary of more than one programme. 67. There may be potential to make greater use of an on-demand approach to targeting in which potential beneficiaries could apply to be considered at specified registration points during a specific window of time. Experience from the HSNP39 indicates that this approach is no more or less effective than the door-to-door registration approach40. Policymakers may want to explore the use of an on- demand approach in future programmes as this would reduce the workload of local implementers and increase their efficiency without reducing targeting accuracy. 68. There are also other opportunities for programmes to learn from each other about the design and implementation of targeting. The ongoing reviews of the OPCT, the PWSD-CT, and the UFS-CT should be looking at lessons learned from the CT-OVC and the HSNP in order to learn lessons from the challenges that they have experienced and to adopt good practices. The CT-OVC should also review the design of the PMT used in the HSNP. Among the key differences between the two PMT designs is that the HSNP uses locally relevant indicators and uses country/district-level poverty rates as a modifier (as opposed to the provincial modifier 41 used by the CT-OVC PMT). These key differences increase the accuracy of the PMT and are therefore improvements that the managers of the CT-OVC may wish to consider making to the design of their PMT.   3.2.2 CHANGE MANAGEMENT AND RECERTIFICATION 69. Beneficiary information can quickly become out of date. Households change location or their poverty status can change, and household members may die or reach adulthood. Also, there may have been errors in documenting names or ID numbers. Analyses looking at household movements in and out of poverty between 1997 and 2007 found that in rural parts of Kenya around 10 percent of the population were newly poor while more than 25 percent had risen out of poverty and become non- poor. The variation was much higher in informal settlements; of the small minority who had been non-poor in 2003, half were poor in 2006.42 It is therefore essential that cash transfer programmes have procedures in place to ensure that the information recorded on the beneficiary lists is kept up to date and that there are systems in place to regularly review households’ continued eligibility. 39 OPM and IDS (2011) 40 It should be noted that HSNP has higher coverage in its target areas than other programme, which may have influenced the effectiveness of the different approaches used. 41 There is a significant weakness in the use of a provincial modifier in Kenya. Most provinces run north to south and therefore include large variations in agro-ecology, with northern areas being drier and poorer than southern areas. 42 World Bank (2008) 18 3.2.2.1 Change Management 70. Most of the cash transfer programmes under the NSNP specify events that will cause households to lose their eligibility such as the death of the eligible household member or the failure to collect three successive transfers. In the event of a household exiting the programme for any such reason, a new household is then selected from a waiting list to enter the programme. 71. Currently, much of the reporting from the district level for the four programmes managed by Ministry of Gender, Children, and Social Development (MGCSD)43 involves this kind of “change management.” This includes correcting beneficiaries’ names and identity numbers as well as dealing with requests to remove beneficiaries from the programme because they are deceased or because they are already benefitting from other programmes. There are similar reporting processes in place for the HSNP. There is no reliable way for district staff to capture such changes as there are no tools in place to capture this information, instead staff are dependent on local information to report deaths or migration. The current systems for documenting such changes are laborious and prone to errors as the reports are typed up at the district level and the information is then entered in the computer at the national level. Reporting is also unlikely to be complete as not all districts submit such reports on a monthly basis. Despite the significant effort expended on this, discrepancies remain as highlighted by a review of the CT-OVC payroll and MIS. This report found a number of discrepancies such as 0.09 percent of transactions involve beneficiaries with the same ID number but different names, 1.98 percent of transactions appear to involve payments to beneficiaries older than 18 years old and a high number (of duplicate names on the payrolls (this can be the result of people sharing the same name, but may also indicate that some households are receiving duplicate payments).44 72. There are also regular attempts to review payrolls to identify any households that have failed to collect three consecutive transfers and to remove them from the payroll (sometimes after checking with the household to ascertain their reasons for not collecting transfers). However, this process is currently done manually for the PWSD-CT, the OPCT and the UFS-CT (and is therefore prone to errors) and tends to be implemented on an ad hoc basis (rather than every payment cycle). 73. There are significant weaknesses in the way in which new households are selected to replace exiting households. Although the programmes are supposed to keep waiting lists, these lists either do not exist or become quickly out of date. Any waiting lists that do exist tend to be generated at the time of the initial targeting exercise and therefore do not include any households who may have become eligible after the targeting process has been completed. As a consequence, replacement households are often chosen on the basis of out-of-date information or are chosen in an ad hoc manner by district officers or village chiefs with little to no documentation to justify their selection.45 74. It is expected that the current review and design process will strengthen and clarify the change management process. Furthermore, change management modules are being added to the MIS systems as described in section 3.3 below. Those responsible for this redesign should consider:  Including a “proof of life” requirement within the OPCT and the PWSD-CT such as the need for the eligible member to be seen by an authorised official or using voice biometrics46 (to complement existing monitoring of CT-OVC beneficiary households whose youngest eligible member is approaching 18)  Automating the review of payroll reconciliations to ensure that households who have failed to collect three consecutive transfers can be identified (this is already in place for the CT-OVC). 43 The four programmes managed by the MGCSD are the CT-OVC, OPCT, PWSD-CT and the UFS-CT. 44 MGCSD (2012) 45 Kimetrica (2013) 46 The technology exists to enable beneficiaries to telephone a proof of life service and for his or her voice to be compared with a voice print taken during programme registration. 19  Improving the selection of replacement households by recommending ways to keep waiting lists up to date and to register newly eligible households (including subjecting them to the same eligibility checks as used during the initial programme targeting exercise). 3.2.2.2 Recertification 75. In addition to ongoing change management, both the HSNP and CT-OVC operational manuals specify the need for the periodic re-registration of beneficiaries in order to confirm their eligibility including their continued poverty status. The CT-OVC manual states that a recertification exercise should be carried out five years after targeting, while the HSNP requires a retargeting exercise upon the launch of the second phase of the programme. 76. However, no recertification or retargeting has taken place within the CT-OVC. This is in part because the number of districts where the programme is more than five years old is small but is mainly because the programme’s limited capacity has been focused on scaling up the programme to new districts. There has also been no retargeting or recertification of beneficiaries in the other programmes managed by the MGCSD, although the retargeting of HSNP beneficiaries is expected to start over the coming months. 77. Retargeting or recertification is needed at regular but not too frequent intervals. Evidence from programmes in other countries indicates that the programme outcomes are better if households continue to benefit from the programme over several years. An evaluation of the HSNP came to a similar conclusion in the Kenyan context as it found that receiving transfers for a longer than average time when combined with small household size yielded better than average results. 47 However, recertifying households too infrequently can result in a beneficiary list that no longer reflects actual vulnerability. 78. It is therefore vital that retargeting should take place in those districts and locations where it is already due and that a timetable should be drawn up for retargeting in all five of the component programmes of the NSNP. Given the dynamic state of livelihoods, there is a need to increase the frequency of recertification from the five years specified by the CT-OVC operational manual to a minimum of once every three years. 3.3 MIS and Single Registry 79. Management information systems (MIS) are essential to safety net programming because of the roles they play in supporting programme administration and monitoring and evaluation. Safety nets have more of a need than some other types of programme for automated MISs due to the high volume of data generated by these programmes. MIS systems for safety net programmes typically have to be able to manage a registry of beneficiaries, generate payrolls, control for fraud and error, and provide regular monitoring information on key programme indicators. 80. At present, the HSNP and the CT-OVC have functional IT-based MIS systems, but the other three programmes are still using paper and Excel-based systems. A recent scoping study48 found that, while both the HSNP and CT-OVC MISs were fit for their purpose, the CT-OVC MIS uses proprietary software, which means that making adjustments to the system can be expensive and not very timely. Furthermore, the CT-OVC MIS is not yet accessible to county or district-level staff to input data with the result that a large number of data entry clerks have to be employed in Nairobi to input data that district staff have already typed up. This delays the verification and reporting of data and increases the chances that mistakes will be made in transferring the information. 81. The key initiatives being supported under the NSNP will be the development of an MIS for each of the other three programmes (the OPCT, the PWSD, and the UFS-CT), the upgrading of the 47 OPM and IDS, (2012) 48 Chirchir and Kidd, (2011) 20 MIS for the CT-OVC, and the establishment of a single registry. These initiatives are being implemented in the following sequence:  Development and Upgrading of Programme MISs. This will ensure that all five programmes have robust MISs that are capable of effectively supporting the implementation of the National Safety Net Programme and of communicating with the single registry. After existing MIS capacity within programmes was assessed, the current open-source web-based MIS used in the HSNP was selected as being the most appropriate model for this purpose. It is also fully compliant with “Management Information Systems Guidelines for Social Protection Programs in Kenya,” a report that was recently produced by an expert working group comprised of staff and consultants from the government and from development partners. The HSNP MIS is currently being adapted as an off-the-shelf customizable MIS to be used in the CT-OVC, the UFS-CT, the PWSD-CT, and the OPCT. This model could also be used in the future for non-NSNP programmes should they have need of it (the World Food Programme is a key member of the working group that is monitoring the progress of the MIS development and may make use of this model for their programmes).  Development of a Core Single Registry: This will also entail using the same open-source web- based MIS used in the HSNP to develop a core single registry for all of the NSNP programmes. This registry will be housed on servers within the Social Protection Secretariat. The single registry will interact in real time with the programme-level MISs to obtain appropriate beneficiary data from each programme. This will provide the Social Protection Secretariat and, by extension, the National Social Protection Council and other policymakers with accurate and up-to-date information on the overall performance of the NSNP. It will also yield data for the core indicators that will be developed as part of the NSNP for the National Integrated Monitoring and Evaluation System (NIMES) that the government uses to track its progress towards meeting its development goals. This includes indicators such as number of households enrolled in the programme, timeliness of payments, and numbers of complaints and whether they have been resolved.  Connectivity with the government’s Integrated Population Registration System. This will involve connecting the single registry to the government’s Integrated Population Registration System (IPRS), which houses data from the national ID system and from the civil register that contains birth and death information. This will make it possible to assess the eligibility of potential and existing beneficiaries more accurately and will significantly minimize the potential for fraud and corruption within the NSNP by verifying that: (i) the beneficiary exists and is alive and (ii) there is a unique ID associated with each beneficiary. The existence of the single registry will also allow the teams implementing the individual cash transfer programmes to access information on any other benefits that their potential or existing beneficiaries may already be receiving from the other programmes. The formalisation of linkages between Single Registry and IPRS has been initiated between MGCSD and Integrated Population Registration Service (IPRS). IPRS is ready to allow access to their web services. 21 Figure 3: Establishment of Kenya’s Single Registry for Cash Transfers Phase 3 Integrated Population Ministry of State for Immigration Registration System & Registration of Persons Phase 2 Single Registry Social Protection Secretariat Phase 1 Ministry of Gender HSNP CT-OVC UFS-CT PWSD-CT OPCT Ministry of Northern Kenya Source: World Bank (2012) 82. This sequence has been developed after significant consultations with programme implementers and other stakeholders about what they needed from the MIS system. This included the preparation of the above mentioned report “Management Information Systems Guidelines for Social Protection Programs in Kenya,” which laid out the requirements for future MIS systems including the single registry. The single registry that is currently being developed is adhering to these guidelines. 83. It is important to recognize that the quality of information contained in any MIS is only as good as the information put into the system, which in turn is very dependent on the capacity of district staff to collect the relevant information and input it into this system. As noted above and discussed further in section 4 below, the lack of staff and resources to conduct all activities related to cash transfers means that this is widely seen as a constraint to the full and proper implementation of the MISs and the single registry. 84. Finalizing these systems and ensuring their functionality is a key priority. However, it is also important to develop the institutional capacity to manage these systems effectively and to make use of the data generated. This has been a weakness in all of the programmes in the past. There has been little attempt to produce standard reports, much less to use the information stored in MIS systems to explore specific performance issues. Enhancements in ICT capacity have taken place recently within the HSNP and CT-OVC, and it is critical to ensure this capacity exists across all of the NSNP. Further work is needed to increase demand for information by programme managers from the MIS system to inform decisions and improve programme management. 85. In addition, further efforts should be made to put into practice the recommendation in the scoping study that information should be entered into the MIS systems at the district level or below (if PDAs are being used, this information can even be entered directly by district children and district gender officers, or even committee members). As stated above, this would increase efficiency, reduce time lags, and reduce the likelihood of errors. 3.4 Payments 86. A payment system can be considered to be effective if it ensures that beneficiaries receive appropriate, reliable, and accessible payments. This means payments that are of sufficient value to achieve programme objectives, are delivered in a secure way that is accessible to beneficiaries, are made as frequently as is commensurate with these objectives, and are paid at predictable times. 22 3.4.1 GENEROSITY 87. The value of a cash transfer needs to be high enough to achieve the programme’s objectives but low enough not to exceed average local incomes (to avoid disincentives) and to be affordable for programme financiers, particularly the government (to ensure equity in coverage and sustainability). 88. In the programmes that make up the NSNP, two benefit levels are currently used. The HSNP provides a monthly transfer of KSH 1,750, while the programmes managed by the MGCSD provide KSH 2,000 per month. The transfer levels differ because they are calculated on a different basis. The HSNP programme transfer level was set in 2008 at a level equal to 75 percent of the cost of a full- food basket food ration (15 kilograms of cereal, 1.5 kilograms of pulses, and half a litre of oil). The MGCSD transfer (initially formulated for the CT-OVC in 2007) was set at a level that took into account the average incomes of the programme target group, the ratio of the transfer to the poverty line, and average monthly expenditures on health and education. 49 The result was a transfer level of KSH 1,500, which represented around 12 percent of the poverty line and 25 to 30 percent of the income of poor households. As Table 5 below indicates, transfers within the NSNP are in-line with similar programmes elsewhere. Table 5: International Comparison of Monthly Benefits from Cash Transfer Programmes50 89. Since transfer values were initially set in 2007/08, there has been significant consumer and food price inflation. In response, the programmes managed by the MGCSD increased their transfer level from KSH 1,500 to KSH 2,000 in December 2011. Despite this increase, real transfers are still less than the original 2007 amount as demonstrated in Figure 4. 49 Alviar, (2009) 50 Author’s own calculations 23 Figure 4: Purchasing Power of CT-OVC Monthly Transfer, 2007 prices 90. There is a need to introduce a system to regularly review and adjust transfer levels to ensure that their value is maintained regardless of inflation. Most long-term cash transfer programmes in the world have such mechanisms and tend to adjust the transfer annually, with many programmes using a simple inflation indexed approach.51 91. In addition to the need for inflation indexed transfers, evaluations of both the CT-OVC and the HSNP highlight the fact that cash transfers have the greatest impact in terms of reducing poverty and increasing consumption on smaller households. This is because a standard transfer is provided to all households regardless of their size and therefore the per capita value of the transfer is higher for smaller households. Table 5 below shows that CT-OVC transfers had a significantly greater impact on the food, health, and education expenditures of households made up of four members or less than on larger households. Table 6: Effect of Household Size in Determining the Impact of a Cash Transfer Indicator Overall Household Size 1-4 5-7 8+ Change in Monthly consumption per adult equivalent (Ksh) 274 ** 486 *** 200 ns 48 ns Change in Monthly food expenditures per adult equivalent (Ksh) 153 * 282 *** 119 ns -7 ns * Change in Monthly health expenditures per capita (Ksh) 17 22 *** 14 ns 14 ns ns Change in Monthly education expenditures per child (Ksh) 27 52 ** 12 ns 8 ns *** ** * ns P-values for mean-difference test, p< 0.01, P<0.05, p<0.1; not significant. Source: UNICEF (2010) 92. This raises the question of whether it is possible to vary the transfer amount by the size of the recipient household. There are several ways of doing this, including:  Full family targeting: A full family targeting approach provides a transfer for each family member, which means that a family of four receives twice the payment received by a family of two.  Per eligible member targeting: An eligible member approach can be applied to programmes that have a categorical targeting element. For example, in the CT-OVC, a household caring for two orphans will receive double the payment received by a family caring for one orphan.52  Per family member or eligible member supplement to a household level transfer: In this approach, a programme would provide a base transfer for the household as a whole, as well as 51 An annual wage rate review is conducted for Ethiopia’s Productive Safety Net Programme, South Africa’s Child Support Grant, and Chile’s Conditional Cash Transfer Program (Chile Solidario) index benefits to inflation, and Brazil’s Old Age Pension is linked to the minimum wage (MoARD, 2010, Degol Hailu et al, 2008, and Samson et al, 2010). 52 If programmes begin to merge their targeting approaches, this will open up the potential for a household that includes both orphans and elderly members to receive an eligible member transfer for each member who is eligible for any of the categorically targeted programmes. 24 a supplement for each family member (or eligible member). For example, a household of six might receive a KSH 1,000 household transfer and six per member transfers of KSH 250. This would amount to a total transfer of KSH 2,500. 93. The government and its development partners should review options for identifying the mechanism that is the most appropriate, easy, and practical to implement in Kenya. 3.4.2 PAYMENT MECHANISM 94. An essential component of any cash transfer programme is a payment mechanism that delivers regular, reliable transfers in a way that is accessible to programme beneficiaries. Such a system needs to involve minimal costs for both benefit providers and recipients, to be reliable and timely, and to provide sufficient documentation to facilitate an effective audit. The accountability and security of the payment mechanism is particularly critical in the Kenyan context because of the concerns about fraud and corruption. 95. Payment mechanisms have evolved since the launch of cash transfer programmes in Kenya. When the first programme (the CT-OVC) was launched in 2004, payments were made by government staff assisted by community volunteers through the District Treasury. By 2006 the decision was made by the MGCSD to outsource payment provision, and the Postal Corporation of Kenya was contracted to provide the CT-OVC payments. As additional government cash transfer programmes were introduced, they also used a combination of the District Treasury and the Postal Corporation of Kenya (PCK). Because of changes in the financial services industry prior to the 2008 launch of the HSNP and of the low coverage by the PCK in Northern Kenya, HSNP managers decided to explore alternate payment mechanisms and thus they contracted with Equity Bank to deliver payments through a network of payment agents. The CT-OVC is now also exploring the use of alternate financial service providers and has also awarded a contract to Equity Bank to deliver a proportion of their transfers. 96. As a result, two payment service providers are currently contracted by cash transfer programmes within the NSNP ‒ the Postal Corporation of Kenya and Equity Bank. Table 6 below compares these two payment services. Table 7: Number of Beneficiaries by PSP in Each Programme Programme Equity Bank PCK CT-OVC 10,400 139,600 HSNP 69,000 OPCT 35,943 PWSD-CT 14,700 UFS-CT 10,200 TOTAL 79,400 200,443 97. A recent strategic assessment53 reviewed the current use of payment service providers by a range of cash transfer programmes in order to make recommendations about how payment service provision could be improved under the NSNP. While the assessment focused on transfers provided by the cash transfer components of the NSNP, it also looked at approaches used by other programmes such as Concern/Oxfam’s use of Safaricom’s mobile money transfer service (M-PESA) in their urban cash transfer pilot and the use by the World Food Programme of Equity Bank’s magstripe cards (which are simple cards requiring PIN authentication) in their Cash For Assets Pilot. Table 7 below indicates the main differences in the services provided by these different payment service providers. 53 Pulver (2012) 25 Table 8: Key Differences in Service Provided by Payment Service Providers Engaged in Cash Transfers in Kenya Postal Corporation of Kenya Equity (Biometric card) Equity (Magstripe card) Safaricom M-Pesa Mobile Money Transfer Payrolls are transmitted manually, with Payrolls are transmitted electronically. Payrolls are transmitted Payrolls are transmitted electronically. four payrolls printed on each site electronically. involving hundreds of pages. Recipients are able to collect payments Recipients are able to collect payments Recipients are able to collect Recipients are able to collect payments from any from their allocated post office branch. from any Equity Bank payment agent payments from any Equity Bank M-Pesa agent. (and can collect money from different payment agent including ATMs payment agents in different months). and bank branches. This is a special service provided for the Payment agents are part of the core This is part of the mainstream This is part of the mainstream payment cash transfer programmes. infrastructure, but use of biometrics is payment infrastructure. infrastructure. specific to the cash transfer programmes and required investment in special equipment such as biometric readers. The cashier checks the recipients’ Recipients are issued with a smart card Recipients are issued with a Recipients are issued with a Safaricom SIM card national ID against the printed payroll with their name and photo on it and with magstripe card and a PIN number. and an M-Pesa account. Using the M-Pesa menu then makes the recipients either sign or biometric fingerprints in the chip. The The payment agent inserts this on their phone, they enter the agent’s number, print their fingerprints on the four copies recipient presents this card at the card into an online card reader (or the amount that they wish to withdraw, and their of the payroll before giving them the full paypoint where the agent inserts it into a ATM) and the beneficiaries then PIN. Both the recipient and the M-Pesa agent amount of their grant. Point of Sale device to authenticate the enter their PIN numbers. receive an SMS text message confirming the identity of the recipient before issuing transaction, and the agent then hands over the the payment. cash amount and requires the recipient to sign a log-book. One-factor authentication (the lack of a Two-factor authentication (biometric Two-factor authentication Two-factor authentication (SIM card and PIN). mechanism to check fingerprints means card and fingerprint). (magstripe card and PIN). that the ID card is the only form of authentication).54 471 branches. 5,300 agents. 149 branches, 560 ATMs, and 45,000 agents. 5,300 agents. 54 There are three possible factors of authentication 1. Something you have (e.g. an ID card or bank card) 2. Something you know (a PIN number or password) and 3. Something you are or do (a biometric finger print or a signature). Two-factor authentication is now the industry standard for money transfers. Two factors cannot be from the same category (e.g. an ID card and a programme ID card). 26 98. Kenya’s regulatory and legislative framework affects social cash transfers because it determines how payments can be made. The Anti-Money Laundering Act, which was passed in 2009, requires that: (i) payment service providers must be regulated financial service providers;55 (ii) all payments must be made in real time online;56 (iii) banks must fulfil stringent “know your customer” checks, thus making it almost impossible to provide payments to recipients without a national ID; (iv) agents cannot be exclusive to a single bank; and (v) no systems can be used that are based on technologies or standards that cannot be integrated with other elements of the national payment system. Although this legislation was passed over three years ago, there is still no timetable for when it will come into force. 99. The strategic assessment 57 made a number of key recommendations that should be incorporated into the NSNP. These include:  New contracts should only be signed with service providers that can comply with the Anti- Money Laundering Act to ensure that programmes are compliant when the legislation comes into force.  Use should be made of the mainstream payment infrastructure. Special purpose infrastructure tends to be very costly and is unlikely to be need given the recent strengthening of the mainstream infrastructure. Furthermore, the mainstream infrastructure must comply with credible international standards, whereas if non-standard approaches are used, this can often leave programme beneficiaries outside of the financial services mainstream. The de facto standard for national and international payments is two-factor authentication using a magstripe card and a PIN.  Beneficiaries should be provided with an account that will allow them to take advantage of financial services, for example, to save a proportion of their transfer. However, this will require more stringent change management mechanisms to monitor that beneficiaries are indeed collecting their payments on a regular basis because it will no longer be possible to easily identify beneficiaries who have not collected three payments in a row. 100. There is also a need to address the current use of short-term contracts with payment providers. While at present these are renewed regularly, they carry the risk of interruption in payment services. 101. Moreover, it is important to ensure that any change in payment provider makes the payment system more accessible to programme beneficiaries. At present, pay points are often situated far from the beneficiaries’ homes and they often have to endure long waiting times at those pay points.58 This can incur both considerable time and financial costs. A recent review of the OPCT and the PWSD- CT found that, in rural areas, beneficiaries frequently spent between KSH300 and 500 on transport costs and travelled on average one or two hours. Furthermore, once at the payment point beneficiaries would also frequently have to wait, sometimes for more than three hours. Selecting payment providers who can offer a larger number of pay points, particularly in rural areas (through a network of agents), should improve this situation.   3.4.3 FREQUENCY, TIMELINESS, AND PREDICTABILITY 102. Evidence shows that the frequency with which payments are made within a cash transfer programme has implications for what type of impacts are experienced by beneficiaries.. Making payments more frequently leads to greater consumption-smoothing by households, while less frequent, lump-sum payments contribute more to productive investments (including the payment of school fees). The frequency of payments can also have cost implications. Small, more frequent 55 PCK is not currently a regulated provider. 56 This will mean that other communication infrastructure (such as satellite-based data services) will be needed to deliver payments in remote areas of the country that lack mobile coverage. 57 Pulver (2012) 58 OPM, (2012a) and Kimetrica (2013) 27 payments result in increased overhead costs, which are not only borne by the programme but also by the beneficiaries who incur time and travel costs in collecting their transfer payments. 103. The timeliness and predictability of transfers can also have a significant impact on a programme’s outcomes. If transfers are not paid on time, this undermines their consumption- smoothing effect and may cause the beneficiaries to go into debt to cover their expenditures until the transfer become available. If transfers are paid on an unpredictable schedule, this limits beneficiaries’ ability to plan their expenditure and make best use of multiple income sources. It may also result in beneficiaries incurring the time and financial costs involved in making repeated visits to their payment service providers to find out if the transfers have been made. For example, an evaluation of the Productive Safety Net Programme in Ethiopia found that the programme’s impact was negatively affected when households received low or irregular transfers.59 104. All of the programmes within the NSNP aim to pay transfers to beneficiaries every two months. Evidence suggests that most of the transfers are being spent on food, followed by health care and education. Therefore, they are succeeding in smoothing the consumption of the beneficiary households in line with the programme’s objectives.60 Given that increasing the frequency with which these transfers are paid is likely to raise costs, maintaining the current frequency seems appropriate. 105. As Figure 5 below demonstrates, in practice, the majority of transfers made by the programmes within the National Safety Net Programme are delayed, thus inconveniencing the beneficiaries and probably reducing the programme’s impact. Delayed payments were the most common complaint amongst beneficiaries in the PWSD-CT and the OPCT61. Data are only available for the programmes managed by the MGCSD, but delayed payments are also reported to be a problem in the HSNP. Figure 5: Timeliness of Transfers Paid by Programmes Managed by the MGCSD, October 2011-September 2012 106. A major reason for the delays in making payments in all of the programmes is the time needed to perform reconciliations at the end of each payment cycle. This is due to the large amount of data that needs to be reconciled and the lack of an automated system for processing reconciliations. For example, in the OPCT a spreadsheet is manually created each cycle to incorporate information from the 300 post offices that pay beneficiaries. In addition, there are sometimes delays in the release of funds by the Treasury to the MGCSD caused either by problems in the transfer request process or 59 Gilligan et al (2009) 60 OPM and IDS, (2012) and OPM (2010) 61 Kimetrica (2013) 28 by a shortage of government funds (see section 5.3 for further details). Furthermore, there has been little monitoring of the timeliness of transfers, which indicates a lack of understanding as to the importance of timely payments. 107. This lack of timeliness and is compounded by poor information flow regarding the timing of transfers. A recent review of the OPCT and the PWSD-CT found that beneficiaries were frequently unaware of the payment schedule and did not know when payments were due as highlighted in the table below: Table 9: Predictability of Payments in for the OPCT and the PWSD-CT OPCT PWSD-CT Nairobi Embu Kilifi Nairobi Embu Kilifi Respondent aware of payment schedule 43.7 25.0 25.4 37.0 25.0 24.1 Payments arrive on time according to schedule 83.1 64.1 73.2 77.8 66.7 72.4 Respondent knows collection dates for next 71.8 56.3 54.9 63.0 61.1 65.5 payment Source: Kimetrica (2013) 108. The timeliness and predictability of transfers needs to be significantly improved under the NSNP. In other countries, this has been achieved by constant monitoring and analysis of any bottlenecks in the transfer process. In Ethiopia, information contained in fortnightly updates on the status of transfers caused the government to prioritize the timeliness of transfer payments, while analysis of bottlenecks revealed the need to keep banks in remote areas better informed to improve their cash flow and to set a schedule for the release of funds at the national level.62 Similar monitoring must become the norm in the NSNP in future, and solutions must be found to the above-mentioned bottlenecks. These solutions might include automating the reconciliations (including linking the payment provider systems with the MISs of the cash transfer programmes) and re-designating transfers in the government accounts to ensure they are accorded the same priority as salary payments. Salaries and emoluments are less prone to delays than other expenditure items in the Government of Kenya’s financial system as they are paid before any other expenditures are made (this is discussed in further detail in section 5.3). There also needs to be improvement in the communication flow regarding payment schedules to beneficiaries.   3.5 Accountability and Grievance Procedures 109. Complaints and appeals procedures are crucial for ensuring the transparency and accountability of cash transfer programmes. These procedures are particularly necessary in relation to targeting and the receipt of payments. Most complaints and grievance systems are expected to fulfil four key functions:  To provide an effective way to resolve any problems that beneficiaries may have with the programme and thus ensure that they receive their full programme entitlements  To collect information that will enable policymakers to make improvements to systems and service delivery (including highlighting serious breaches in implementation such as fraud or abuse).  To enable programme beneficiaries to influence programme design, management, and service delivery.  To build trust in and ownership of the programmes by citizens, thereby making the programmes more sustainable. 110. There are a number of key principles and characteristics of a well-functioning complaints and appeals system:  Knowledge of programme entitlements and complaints mechanisms is widespread among the target population. 62 Based on the authors’ personal knowledge 29  Beneficiaries have easy access to complaint mechanisms, including being able to make a complaint to someone other than the person that the complainant feels is responsible for the grievance, and a belief that making a justified complaint will result in change.  The system has the authority to make binding decisions.  The system gives feedback to the complainant about whether and how a complaint has been resolved. 111. The majority of the programmes that make up the NSNP include some form of complaints and grievance procedure on paper. In addition the 2008 service charter of the MGCSD outlined a series of commitments by the MGCSD along with a complaints mechanism that people can use if they feel that these commitments have not been fulfilled. However, only the CT-OVC and HSNP have fully functioning mechanisms including community oversight committees, external grievance and appeals mechanisms, and beneficiary service charters. 112. Despite the significant progress that has been made in establishing complaints mechanisms in both the HSNP and CT-OVC, further work is needed in terms of the case management aspect of the complaints. Recent reports on the HSNP, the OPCT and the PWSD-CT highlight beneficiaries’ perceptions that any problems that they raise are not resolved63. Meanwhile, there is some evidence that complaints made about the CT-OVC are also rarely resolved. For example, those who complain about being excluded from the programme are told that they will be considered in any future retargeting process, while those complaining about late payments are informed of a revised payment date.64 Therefore, although feedback is provided no action is undertaken to resolve the cause of the complaints. 113. There are also very significant differences in the numbers of complaints recorded in the different programmes. These differences highlight the varying levels of knowledge about the complaints mechanisms and their differing levels of functionality. One indication of the lack of functionality in the OPCT and the PWSD-CT is that the majority of complaints were never officially logged, even when they were reported to officials. Even in the HSNP, which has the highest level of complaints, beneficiaries’ knowledge and understanding of programme entitlements and complaints mechanisms is still limited, and in other programmes it is even lower (see Table 10 below). Therefore, all programmes need to make greater efforts and investments in spreading information about their entitlements and complaints mechanisms. Table 10: Awareness of Complaints and Grievance Mechanisms by Beneficiaries of Each Cash Transfer Programme CT-OVC HSNP PWSD-CT OPCT UFS-CT Between 10 and Only 15% of 15% of 15% of No data are 47% of respondents feel beneficiaries beneficiaries available. beneficiaries are that they would understand how to understand how to aware that they know who to talk make a complaint. make a complaint. can make a to if they had a c/ c/ complaint. a/ problem with the programme. b/ Sources: a/ Kimetrica (2012c). b/ OPM (2012a). c/ Kimetrica (2013) 114. The World Bank is currently providing technical support to the government to review the complaints and grievance mechanisms of programmes managed by the MGCSD. It is expected that this will result in significant improvements to the design of these mechanisms. Furthermore, the development of the grievance module of the upgraded MIS that will be used by all programmes will make it easier to ensure consistent recording and analysis of grievances across all five programmes 63 OPM (2012a) and Kimetrica (2013) 64 Internal communication from Kimetrica 30 whether such grievances are registered through grievance mechanisms managed by implementers (such as the MGCSD) or contracted out to service providers (as is currently the case for the HSNP and in-part the CT-OVC). 115. In addition to the above changes, the technical assessment has identified the following areas that could further improve the design and functioning of grievance procedures:  As households will continue to make complaints to the authorities that they know (such as targeting committee members and location chiefs), it is important that these people retain some role in documenting and forwarding complaints through the defined mechanism.  The experience of beneficiary committees (the BWC under the CT-OVC and the Rights Committee formed for the HSNP) seems largely positive. Similar committees, with elected volunteer beneficiaries trained to receive and handle complaints, should be considered for the remaining three programmes.  Further work needs to be done to ensure as many targeting-related appeals as possible are dealt with during the period when targeting decisions can easily be revisited. For this to be achieved, households must be made aware of their rights to appeal prior to the closing of the targeting window. It will also be necessary to allow a sufficient amount of time between the posting of the initial list of beneficiaries and its final validation as described in section 3.2.1.2. It is during this period that appeals can most easily be considered and resolved.  There may need to be further work to ensure that any complaints made to sub-contracted service providers are also recorded and resolved. Many complaints about late payments, incomplete payments, or problems with the payment process will (appropriately) be made to the payment provider involved. Therefore, it is crucial that the programme managers are aware of these complaints and that any that cannot be resolved by the sub-contractor can be referred to the programme itself.    There needs to be a renewed focus on the effective communication of programme objectives, targeting criteria, systems, entitlements, and the complaints system so that beneficiaries and community members can effectively hold service providers to account.    The case management of complaints needs to be improved to ensure that any problems experienced by beneficiaries are resolved and that they are notified about whatever the action is being taken.    As the Government of Kenya pursues its agenda of increasing the coordination and harmonization of cash transfer programmes, introducing a common complaints and grievance system across some or all of the cash transfer programmes might be a way to increase efficiency. 116. In addition, the National Social Protection Policy states that independent regulators will regulate, set standards for, and supervise compliance by social security and health insurance schemes. The policy states that one or more institutions will be set up to hear appeals and resolve social protection disputes in addition to any internal complaint mechanisms set up by implementing institutions. When such institutions are established, they will need to become mainstreamed into the grievance procedures of the NSNP.   3.6 Building the Capacity of the NSNP to Scale up Support in Response to Shocks 117. Many Kenyans live in drought-prone environments 65 or are susceptible to floods, price increases or other shocks. FEWS NET (USAID’s Famine Early warning System) has identified the following groups of Kenyans who are particularly vulnerable to shocks:  Pastoralists in the north, northeast, northwest, and the southern Maasai rangelands  Destitute pastoralists who have dropped out of the pastoral economy after losing livestock to droughts, floods, and conflict and who are now located close to urban centres near their traditional grazing areas 65 Approximately 34 percent of the population lives in the vulnerable arid and semi-arid lands where drought is particularly prevalent. 31  Agropastoralists in the south and northwest  Marginal agricultural households in the coastal and southeastern lowlands  Marginal agricultural and fishing households situated around the lakeshore  Urban poor in most of the major towns  Households displaced during the post-election chaos in late 2007 and early 2008.66 118. This vulnerability, combined with longer and more frequent droughts, has led to increasing numbers of people who require emergency food aid. Over the past eight years Kenya has experienced four episodes of severe drought (in 2004/05, 2005/06, 2008/09, and 2010/11). Figure 6 below indicates the numbers of people who have required emergency food aid since 2004. Figure 6: Population Requiring Emergency Assistance March 2004 - February 2013 Source: Data provided by the VAM Unit (Vulnerability Analysis and Mapping Unit) of the World Food Programme in December 2012 119. The new Government of Kenya’s Drought Risk Management and Ending Drought Emergencies Medium-Term Plan 67 identified a number of failures in the current system for addressing droughts and emergencies. These included the fact that responses were often late, there was a failure to respond to existing early warning information, and emergency responses were reactive (a crisis management approach) rather than anticipatory (a preventive risk management approach). 120. There is significant evidence that a faster response is possible when existing instruments are used instead of developing new ones. The UFS-CT was initially planned in 2009 as an emergency response to high food prices and their effects on urban populations but took two years to become operational. By comparison, Ethiopia’s Productive Safety Net Programme (PSNP) includes a risk financing mechanism (RFM) that allows the programme to be scaled up in the event of a shock. The time between the request for PSNP risk financing resources and disbursements was six weeks in 2011. See Box 2 for further details of the speed of the Ethiopian risk financing mechanism response. Box 2: PSNP’s Risk Financing Mechanism 2011 saw a major drought affecting a number of countries across the Horn of Africa, including Ethiopia. A humanitarian appeal was launched in March 2011, five months after the semi-annual seasonal assessment was completed. While the March appeal resulted in some resources being made available for the response, as of December 2011 (nine months after the appeal was launched and some 13 months after the original assessment), 94 percent of the funding for the humanitarian appeal was in place. By contrast, in August 2011, when regular 66 USAID (no date) http://www.fews.net/ml/en/info/Pages/fmwkfactors.aspx?gb=ke&l=en 67 Republic of Kenya (2013) 32 PSNP transfers stopped, the risk financing mechanism (RFM) completed a rapid verification of needs in highland areas within a month of the request for RFM resources, and financing was disbursed within two weeks of the request. From request to disbursement took six weeks. This shows that, when the preconditions are met, the RFM easily outperforms the humanitarian system in terms of verifying needs and disbursing resources for a response to be delivered through government systems. While an assessment is required to determine the impact of the RFM on livelihoods, the RFM’s early and preventive response to an identified need means that it has a far higher chance of helping affected people not to resort to negative coping strategies and asset depletion as a response to a shock. Source: http://www.odihpn.org/humanitarian-exchange-magazine/issue-53/how-ethiopias-productive-safety-net- programme-psnp-is-responding-to-the-current-humanitarian-crisis-in-the-horn 121. Cash transfers tend to be cheaper and faster to deliver than food aid. Food aid often has to be procured internationally, but even when it is procured locally, it can take several months before procurement processes are completed. The delivery of food aid also involves significant costs and takes time, whereas cash transfers sent through the banking system tend to be much faster. The combination of lower costs and faster speed means that the aid has a greater impact in terms of smoothing consumption and protecting livelihoods per dollar spent. 122. There is currently a process underway to design a mechanism for scaling up the National Safety Net Programme in response to shocks. This design process will review the current situation and design the following mechanisms:  The Early Warning System. The current early warning system is being reviewed, including indicators and the phased classification system to identify indicators that will be used to trigger the release of contingent financing.  Contingency Planning. The government has already invested in district/county-level contingency planning tools, but they do not include a cash transfer response. Therefore, it is necessary to adjust the current tools to include a way to increase the coverage of cash transfers or their value. This should include mechanisms for determining the most appropriate scale of support, how resources will be targeted to affected households and how scaled-up cash transfers will work alongside other emergency support or long-term food or cash transfers.  Contingent Financing. The design process will consider how the contingent resources, once triggered, will flow from the National Contingency Fund through the cash transfer programmes to the targeted beneficiaries.  Monitoring and Evaluation. This will establish how programme managers will monitor the scaling up of the National Safety Net Programme within the existing monitoring systems for the cash transfer programmes and the National Drought Management Agency (NDMA). It will also consider how the government will assess the effectiveness of this overall approach. 123. This approach will be housed within the government’s existing National Drought Management Authority (NDMA), will be consistent with the government’s Drought Risk Management and Ending Drought Emergencies Medium-Term Plan,68 and will take advantage of the established early warning system and local contingency plans enhancing ownership and improving coordination with other humanitarian interventions. Financing will flow through a National Drought and Disaster Contingency Fund to be established under the Ministry of Finance with a Board of Trustees responsible for facilitating rapid but transparent and accountable release of funds. The decision to incorporate a mechanism to scale up transfers in response to shocks addresses these findings and ensures that the NSNP is among the few programs involved in best practice in this area. 124. It will be critical to incorporate lessons learned from other countries. For example, the Ethiopian experience has highlighted the potential effectiveness of a risk financing approach but has also identified challenges: 68 Republic of Kenya (2013) 33  Reponses financed through the risk financing mechanism are managed and considered separately from other humanitarian responses. This artificial divide results in misunderstandings about who is covered, poor information exchange between the actors implementing a humanitarian response, and the managers of the PSNP, and a lack of coordination between the different responses.  There is a disconnect between the mainstream needs assessment process for the humanitarian response in Ethiopia and how the risk financing mechanism is triggered and allocated. It is critical to ensure that there is investment in a common early warning system which is sufficiently robust and appropriate to inform all humanitarian responses whether within or outside of the NSNP.  Despite investments in the development of manuals and guidance materials, there continues to be confusion about the appropriate sequencing of different humanitarian response tools. As a result, the risk financing mechanism is often not used to trigger an early response but is instead used a way to fill any gaps left by an inadequate response to the humanitarian appeal.69 125. In addition, there is a need to work with payment service providers to develop a secure payment mechanism that can be put into action immediately if and when new beneficiaries are identified and that will be flexible enough to allow the paying of different transfer rates. This might involve the use of pre-paid payment cards or the issuing to beneficiaries of magstripe cards. 3.7 Complementary Actions and Conditions 126. Many cash transfer programmes seek to enhance their impact by providing complementary programming or by putting certain conditions on the receipt of benefits, such as requiring recipients to make use of health services or to keep their children in school. The NSNP intends to implement two key complementary activities ‒ birth registration and use of national ID cards and the piloting of conditional cash transfers within the CT-OVC – covered in sections 3.7.1 and 3.7.2. A third section highlights other complementary interventions implemented outside the NSNP. 3.7.1 BIRTH REGISTRATION AND NATIONAL ID CARDS 127. Having a birth certificate and/or a national identity card can enable individuals to access critical services such as health care, education, and bank accounts. Two of the components of the NSNP focus on improving how beneficiaries can access such documentation. 128. One objective of the CT-OVC is to increase the number of OVCs who have birth certificates and, with this in mind, responsibility for encouraging households to register their children’s births was given to the Location OVC committees. The 2010 evaluation of the programme found that the proportion of children from beneficiary households with birth certificates or birth registration forms had increased by 12 percentage points over the control group during the period 2007 to 2009 and that 92 percent of caregivers were found to hold national identity cards by 2009 (compared to 86 in 2007).70 129. The decision to use the banking sector to deliver HSNP cash transfers highlighted the need for beneficiaries to carry national identity cards. New legislation (described above in section 3.4.2) has required banks to “know your customer” and this has made having a national ID a requirement for HSNP beneficiaries to receive their cash transfers from a payment service provider. In the first year of the programme’s operation, the number of people holding national ID cards has not changed despite the increased incentive to register. This may reflect the limited capacity of government agencies in the programme districts which meant they were unable to increase civil registration. Those beneficiary households who had no member with a national identity card were forced to nominate a neighbour as 69 Hobson and Campbell (2012) and Sandford (2012) 70 OPM (2010) 34 a payment recipient, but this carries the risk that the payment recipient will not pass on the full transfer to the intended beneficiary. 130. The availability of national ID cards also has the potential to reduce opportunities for fraud and error in the programme. The use of a unique ID by each beneficiary will make it possible to check all programme MISs to ensure that households are only receiving one cash transfer. It will also allow early detection of any errors in the transcription of people’s details and will reduce the potential for including ghost beneficiaries. However, a recent Payroll and MIS Technical Review found that many programme beneficiaries (representing 9.02 percent of transactions) continued to use temporary, programme identity cards to collect payments beyond the three cycles allowed by programme regulations. There are also incidents, few in number (representing only 0.09 percent of transactions), of duplicate national identity card numbers being use to collect more than one payment in a given payment cycle. 71 131. Any future support to the sector should strengthen the focus on encouraging adults in beneficiary households to register for national ID cards across all five programmes. In addition, new legislation regarding payment service providers (described above in section 3.4.2) will require all recipients of transfers to hold national ID cards. Those beneficiaries without a National ID will have to rely on a nominated payment recipient with the attendant inconvenience and risks associated with such an arrangement. In addition, the programmes making up the NSNP should make better use of the opportunity this provides to conduct payroll checks and follow up any discrepancies which emerge.   3.7.2 PILOTING OF CO-RESPONSIBILITIES WITHIN THE CT-OVC 132. A new pilot is currently being designed to test whether it is possible to enhance the positive impact of cash transfers on the human development outcomes of children by requiring beneficiaries to meet specific conditions or “co-responsibilities” involving the use of health and education services. The pilot builds on lessons learned from a previous pilot that faced significant challenges in ensuring that the beneficiaries understood their co-responsibilities and in effectively monitoring their compliance. The co-responsibilities under the new pilot will include:  Each child aged between 0 to 1 years old must attend a health facility for immunizations every two months.  Each child aged between 1 to 5 years old must attend a health facility for growth monitoring and vitamin A supplementation every six months.  Each child between 4 to 17 years old must attend school and should not be absent from that school for more than three days per month.  The caregiver herself/himself or an adult member of the household must attend awareness sessions at least once in a year or when scheduled by the District Children’s Officer. 133. Beneficiaries’ compliance with their health and education co-responsibilities will monitored at the payment site during the months of April, August, and December each year (with the first monitoring exercise due to take place in April 2013). Beneficiary caregivers of OVC need to bring proof of compliance to the payment site, including a school report form and a Ministry of Health form 216. If one household member fails to comply with any of the programme’s conditions within the compliance monitoring period, KSH 500 will be deducted from the next payment disbursed to the household (this is increased to KSH 1,000 if two household members fail to comply). 134. International experience has shown that social protection programmes that provide conditional cash transfers can significantly increase the human capital of an extremely poor population in the medium term while reducing the poverty gap in the short term.72 However, there is debate as to whether the human development impact is significantly influenced by the conditions or is 71 MGCSD (2012) 72 Fiszbein and Schady (2008) 35 simply due to the cash transfers themselves as unconditional cash transfers have also been proved to build human capital. There is growing evidence that, while conditional and unconditional transfers have both significant positive effects, the impact of the two methodologies varies. A study undertaken in Malawi with the purpose of assessing these differing effects found that conditional transfers had reduced the school dropout rate much more significantly than unconditional transfers but had less of an impact on other important indicators such as early marriage or pregnancy.73 135. However, it should be noted that for these conditions to be effective in achieving the programme’s objectives of improving the welfare and increasing the resilience of beneficiaries, then it is essential for the other key services – such a schools and health clinics – to be available and adequate to support the poverty-reducing effect of cash transfers. 136. The launch of the new pilot programme has been subject to delays because of capacity restraints and the challenges involved in introducing new approaches. When monitoring and evaluating the conditional pilot, policymakers should compare outcomes in the areas where conditional transfers are being piloted with outcomes in the areas where only unconditional cash transfers are offered. They should also consider the practicality of introducing conditional transfers on a large scale given the capacity constraints that exist in Kenya. These capacity constraint have implications for both the provision of adequate key services and the ability to monitor adherence to conditions.   3.7.3 COMPLEMENTARY INTERVENTIONS TO SUPPORT POVERTY REDUCTION 137. The National Social Protection Policy emphasizes the need for programmes to include appropriate graduation strategies to help households to “wean themselves off social assistance schemes and programmes and become financially self-sufficient wherever possible.” The policy goes on to describe the importance of skills development and access to micro-finance services and other start-up capital to enable beneficiaries to rise out of poverty. Providing this kind of livelihood- enhancing support is not the only way to help households to graduate from cash transfer programmes. Other key ways to complement the positive impact of cash transfers include increasing access to health and education services, creating employment opportunities, and funding interventions that reduce the marginalization of some populations. 138. The Ministry of Gender, Children, and Social Development (MGCSD) and the Ministry of State for the Development of Northern Kenya and Other Arid Lands (MSDNKAL) already offer a range of development activities that could complement cash transfers and promote the graduation of beneficiary households from poverty. These include:  The piloting of conditional cash transfers as described above.  The Women’s Enterprise Fund, which provides alternative financial services to women excluded from the mainstream formal and informal financial sectors. The fund provides affordable credit to help women to start or expand small-scale businesses.  The National Development Fund for Persons with Disabilities, which includes components supporting assistive devices and services, educational assistance, and revolving funds for groups of people with disabilities.  Activities and projects focused on orphans and vulnerable children including interventions that increase the access of OVC to essential services such as education and health as well as child protection and psychosocial support to OVCs and other children through District Children’s Officers. The potential for linking the targeting of the CT-OVC and the National Hospital Insurance Fund indigent scheme are also currently being explored.  A range of drought management and community development initiatives in the arid lands of Northern Kenya including the Arid Lands Resource Management Project and the Kenya Drylands Livestock Development Programme as well as plans for activities such as livestock insurance to complement the HSNP. 73 Baird et al (2011) 36  Livelihood-enhancing activities for households living in informal settlements, which are currently being piloted by Oxfam and Concern with government interest in their scale-up. 139. There is considerable scope for increasing coordination between the cash transfer programmes and other interventions implemented by the MGCSD, the MSDNKAL, and other actors. This coordination needs to be increased both at the ministerial level and within the districts where programmes are operational. In addition to ensuring the coordination of all of the government’s own programmes, it is crucial to increase knowledge of the cash transfer initiatives so that other relevant agencies, such as NGOs, can ensure that their interventions complement and are coordinated with those of the government. 140. While greater coordination is crucial for increasing the impact of cash transfers in terms of reducing poverty, it is important to be realistic about how much graduation can be expected. Kenya’s resource constraints mean that cash transfers will only be targeted to the extreme poor who will require support over long periods of time before there is any likelihood of significant numbers of them becoming financially self sufficient. Furthermore, the additional vulnerability experienced by much of the NSNP target group (including old age and severe disability) limit the extent to which they can engage in livelihood-enhancing activities. 3.8 Gender 141. Cash can empower poor individuals by enabling them to make decisions to improve their lives and the lives of their families. Cash transfers can also significantly increase the empowerment of women within the household and community as well as resulting in other important effects such as increased girl’s education and greater access to and use of health care. Programmes can enhance this positive impact by taking into account the greater poverty of female-headed households, making special provision for women to participate in programme decision-making, and making female caregivers the primary recipient of transfers and complementary activities. Table 11 below shows how gender considerations have been incorporated into the design of the five cash transfer programmes of the NSNP. Blank cells indicate either that this aspect has not been incorporated into design or that there is no available information on how or whether it has been considered. As this table shows, further work is needed to ensure that gender considerations are better addressed in the OPCT, the PWSD-CT and the UFS-CT and the currently ongoing reviews and manual preparation will help to address this. Table 11: How Gender has been Incorporated into Aspects of Programme Design CT-OVC HSNP OPCT PWSD-CT UFS-CT Vulnerability of female-headed   households recognized in programme targeting Special provision for women to * # participate in programme decision-making Women selected as primary # # recipients of transfer Availability of gender   disaggregated data Notes:  - Indicates that there is information that the issue has been addressed. # - Indicates that although there is no information, there is some evidence that practice is in line with the indicator. * - Indicates that the Location OVC committees, which conduct targeting, make no specific provision for women but that provision is made in the Beneficiary Welfare Committees. Guidance states that no more than 2/3rds of their members should be from one gender. 37 142. Evaluations of the HSNP and the CT-OVC have shown that, although women are not explicitly targeted as the main recipient of transfers, in practice the majority of recipients are women74 (70 percent in the HSNP and 92 percent in the CT-OVC). 75 Both evaluations indicated that the transfer recipient plays a significant role in deciding how the cash transfers are spent (for example, the person who normally decides how the HSNP cash is spent is female in 63 percent of beneficiary households). The HSNP evaluation indicated that there may be small spill-over effects into household decision-making in general. These benefits come with potential costs. Women incur the significant time-costs related to collecting transfers, including the need for repeated visits to payment service providers as a consequence of delayed and unpredictable payments. Improving the timeliness and predictability of payments and increasing the accessibility of payment agents will help to mitigate these costs. 143. While cash transfer programmes in other countries have often had other gender-related effects, the picture is more mixed in Kenya. Cash transfers do appear to have increased enrolment in secondary schools, but this impact is only statistically significant for boys. 76 An analysis of the impact of the CT-OVC on the transition to adulthood77 found evidence that previous participation in the CT-OVC reduced risky sexual behaviour by young adults, (particularly young women) and had a positive impact on their mental health (particularly for young men). 144. Policymakers should ensure that gender is more effectively taken into account in the design and processes of the NSNP. It is particularly important for women to be enabled to participate in programme decision-making and for programme data (on both programme participation and impact) to be disaggregated by gender. Having more gender-disaggregated data will help programme designers to reform the NSNP to improve its gender-related outcomes. The draft results framework for the NSNP programme requires gender disaggregated reporting against key indicators and the new MISs should make it easier to collate and analyse gender-disaggregated data and facilitate monitoring and reporting. This information will be combined at the programme level in the single registry and key indicators reported to the government’s central Monitoring and Evaluation Directorate for inclusion in the National Integrated Monitoring and Evaluation System (NIMES). 4 INSTITUTIONAL ARRANGEMENTS 145. This section describes and assesses the existing institutional architecture of the NSNP and the adjustments that are expected to be made to this architecture. It does so by considering the institutional set-up at each level of government - the national, district, and community levels. It also looks at the incentives that will affect programme performance and programme mechanisms for the management of fraud, error, and corruption. 4.1 Description of Existing Institutional Architecture 4.1.1 OVERARCHING NATIONAL LEVEL MANAGEMENT 146. The National Council for Social Protection, chaired by the MGCSD, provides overall strategic direction for social protection in Kenya. This committee is supported by a Social Protection Secretariat, which is currently located within the Department of Gender and Social Development within the MGCSD.78 This body is currently weak as its role is still evolving. Its initial focus was on developing the National Social Protection Policy, and is currently reviewing its functions and filling 74 It is interesting to note that the breakdown of beneficiary households which are male and female headed are nearly equal. 75 OPM (2010) and OPM and IDS (2012) 76 OPM (2010) 77 Handa et al (2012) 78 It is expected that the secretariat will move out of this department but will remain within the MGCSD once the management structure for social protection has been fully developed and put into operation. This management structure is described in more detail in section 4.2.1.1 below. 38 staff positions. The secretariat is expected to play an increasingly important role in the coordination and consolidation of the different cash transfer programmes, but at present there is a lack of clarity regarding how it will work with the units that are responsible for day-to-day management of these programmes. 147. The five cash transfer programmes that make up the Kenya National Safety Net Programme are currently administered by two different ministries and three different departments. As outlined in Figure 7 below, the HSNP is housed in the Hunger Safety Net Secretariat under the Ministry of State for the Development of Northern Kenya and Other Arid Lands (MSDNKAL); the CT-OVC is housed in the Department of Children’s Services within the Ministry of Gender, Children, and Social Development (MGCSD); and the remaining three cash transfer programmes (the UFS-CT, the OPCT, and the PWSD-CT) are supported by the Department of Gender and Social Development also within the MGCSD. Figure 7: Current Institutional Structure for Administering Cash Transfer Programmes Cash Transfer Programs Ministry of State for  Ministry of Gender, Children, and Social Development Development of Northern  Kenya and Other Arid  Lands Department of  Department of Gender and Social  Hunger Safety Net  Children's Services Development Secretariat CT‐OVC UFSP OPCT PWSD‐CT HSNP   148. The institutional arrangements are a reflection of how programmes have evolved and the divisions between departments and ministries. As a consequence, the structure of the HSNP varies considerably from that of the other four programmes while differences within the programmes managed by the MGCSD are partly due to the differing sizes and maturity of the programmes. The different institutional arrangements of the programmes are described in more detail in the sections below. 4.1.2 INSTITUTIONAL ARRANGEMENTS OF MGCSD-MANAGED PROGRAMMES 149. The programmes managed by the MGCSD follow the general government administrative structure with national-level management and oversight, a small amount of monitoring and supervision by province-level staff, district-level coordination of implementation and village, and community-level involvement. 4.1.2.1 CT-OVC 150. In addition to the above national oversight mechanisms, the CT-OVC has its own National Steering Committee that oversees and approves policy developments on orphans and vulnerable children. The Steering Committee is made up of relevant ministries and departments of the Government of Kenya, development partners, international NGOs, faith-based and civil society organizations, and the private sector. 151. The Department of Children’s Services acts as the secretariat for the OVC Steering Committee while a central programme unit (CPU) within the secretariat is responsible for managing the day-to-day implementation of the CT-OVC. The CPU is responsible for preparing annual work plans and budgets, producing quarterly and annual progress reports, organizing mid-term reviews, and 39 managing the procurement and financial management functions (including the management of payment service providers). 152. At the district level, District Children’s Officers (DCO) coordinate local implementation and play a lead role in the programme’s outreach activities, coordinating the targeting process, and supporting families through case management. The DCO is supported by the Area Advisory Council, which supports all children’s services and the District OVC Sub-Committee (DOSC), which is focused on the CT-OVC. The DOSCs support the DCO in conducting programme outreach, in the formation, identification, and training of Location OVC Committees and enumerators, and in the oversight of targeting and other programme activities. 153. Payments are the responsibility of outsourced payment service providers. At present the Postal Corporation of Kenya and Equity Bank are sub-contracted directly by the MGCSD to provide payment services. 154. At the community level, unpaid Location OVC Committees (LOCs) play a key role in raising awareness of the programme among community members and carry out the first stage of targeting prior to locally employed enumerators collecting more detailed information from potential cash transfer beneficiaries. LOCs also organize community barazas, which are held at key stages of programme implementation, for example, during the community validation of beneficiaries stage. Beneficiary welfare committees (BWCs) are in the process of being established in operational areas. These committees, composed of and elected by beneficiaries, are responsible for empowering and representing beneficiaries by, for example, informing beneficiaries of payment dates, providing a list of beneficiaries who did not receive a payment during a payment cycle to DCOS to facilitate follow- up, collecting and forwarding information to DCOs concerning any changes in beneficiary status, and forwarding any complaints submitted by beneficiaries to DCOs.   4.1.2.2 UFS-CT, OPCT, and PWSD-CT 155. The institutional arrangements for the OPCT, the PWSD-CT, and the UFS-CT are similar to those of the CT-OVC. At the national level, responsibility for implementing these programmes lies with the Division of Social Welfare of the Department of Gender and Social Development within the MGCSD. At the district level, District Officers identify programme beneficiaries in consultation with relevant community leaders, committees (where they exist), and – in the case of the UFS-CT – enumerators. The payment process is contracted out to the Postal Corporation of Kenya. Community committees and/or Chiefs, Sub-chiefs, and village elders are involved in targeting and programme implementation. 156. In addition to this basic structure, the National Council for People with Disabilities also plays a role in the PWSD-CT. The National Development Fund for Persons with Disabilities, which is the funding arm of the council, finances the cash transfers, and the National Council oversees implementation by the MGCSD. The council is also expected to enhance coordination with complementary activities such as the other grants provided by the development fund. 157. There has been less investment in systems within these three transfer programmes than in the CT-OVC and HSNP programmes. This is partly a result of their smaller size and their relative newness, but it is also due to a lack of technical assistance from development partners. In particular, programme documentation is weak and there is no comprehensive management information system. As already highlighted in section 3 above, work is already underway to review the design of these programmes and to document procedures in comprehensive manuals. MIS systems are also being developed in parallel with the development of a single registry.   4.1.3 HSNP 158. At the national level, the HSNP is the responsibility of the HSNP Secretariat within the Ministry of State for Development of Northern Kenya and other Arid Lands (MSDNKAL). Key 40 functions of the HSNP Secretariat include preparing work plans, establishing mechanisms of coordination, regularly reviewing programme progress and managing fiduciary risk. At the national level, DFID, the main programme funder to date, has played a key role in programme oversight and in contracting with the lead agencies responsible for implementing different components of the programme. Although the secretariat is based within MSDNKAL, the Government does not provide any financing for the programme and HSNP staff are short-term contractors financed by the development partner. This undermines the sustainability of the structures, with the consequent risk that programme systems and procedures, and the staff who operationalize them, may easily be lost. 159. The large distances covered by the programme combined with severe capacity constraints within the government administration has meant that all key areas of the HSNP’s implementation are sub-contracted:  The administrative component, including targeting, enrolment, registration, and case management, is implemented by a consortium of NGOs headed by Oxfam GB.  The payments component is managed by the Financial Sector Deepening Trust,79 which in turn has sub-contracted with Equity Bank to deliver payments to beneficiaries. Equity bank uses a network of agents to administer these payments.  The social protection rights component, which is designed to spread awareness of programme rights, entitlements, and responsibilities and to manage the grievance mechanism, is implemented by HelpAge International.  The monitoring and evaluation component is implemented by the Oxford Policy Management consulting company, which collects and analyses data on a range of process and impact indicators.  The management information system was initially managed by an independent consultant but since 2010 has been managed by an MIS Officer based in the HSNP Secretariat. 160. A number of mechanisms now exist to facilitate coordination among the various agencies that are implementing the HSNP. For example, at the national level, a coordination meeting is held every fortnight to share information on progress and on the resolution of any implementation challenges. One key aim of this forum is to pool information about and learn lessons from each agency’s implementation experience and use it to improve the programme in future. Monthly district-level coordination meetings are also held to share work plans and progress, and these play a key role in resolving implementation issues. This is the key forum in which beneficiaries’ grievances are discussed. 4.2 Anticipated Changes to the Institutional Arrangements 4.2.1 CHANGES RESULTING FROM THE NATIONAL SOCIAL PROTECTION POLICY AND THE LAUNCH OF THE NSNP 161. Figure 8 below lays out the expected institutional architecture of the NSNP. Key adjustments to the current arrangements are highlighted in yellow. The main adjustments include:  A National Social Protection Council (NSPC) will be created to bring together all key stakeholders in the area of social protection (as laid out in the new National Social Protection Policy). It will consist of representatives of the government ministries engaged in social protection, as well as businesses, social security organizations, and CSOs and FBOs.  The council will be supported by a National Social Protection Secretariat that will implement the NSPC’s decisions and to carry out day-to-day functions. The current Social Protection Secretariat (described in section 4.1.1 above) will take on this function. The secretariat will provide the NSPC with technical support and will coordinate the implementation of social protection actions. 79 FSD Kenya is an independent trust established to support the development of inclusive financial markets It works in partnership with the financial services industry to expand the access of lower-income households and small-scale enterprises to services. 41  The Treasury will play a greater role in future cash transfer programming in accordance with the Programme for Results approach. As the programme will be relying on mainstream government financing and will be focused on ensuring that disbursement linked indicators are met, it will be particularly important for the Treasury to be closely involved with programme processes.  The management of the HSNP will become the responsibility of the National Drought Management Authority (NDMA). The NDMA will also have the responsibility for the scaling up of cash transfers in response to shocks. The Authority will oversee the collection of early warning information, monitor whether financing should be triggered, and oversee the disbursement of funds.  The National Drought and Disaster Contingency Fund (NDDCF) will be established as a Trust Fund under the Ministry of Finance. This Fund will finance the scaling up of the NSNP in the event of a shock. The NDDCF will be managed by a Board of Trustees constituted of Permanent Secretaries of the MSDNKAL; the Ministry of Special Programmes or such Ministries charged with disaster responses; Finance; and, at least three other members with expertise on drought management and crisis response.80  Social Protection Committees at the county and sub-county levels will also be formed, as outlined in the National Social Protection Policy. As social protection is a function of the national government, the committees will be answerable to the National Social Protection Council but will establish appropriate reporting relationships with county governments. Their main responsibilities will include overseeing and monitoring social protection interventions, coordinating and harmonizing programmes, and ensuring that sector policies and guidelines are implemented within the county. Figure 8: Organogram of the Key Institutions Involved in NSNP Implementation 80 The Board of Trustees shall be responsible for determining the policy of the Fund and controlling its operations; it shall hold the Fund in trust for the National Drought Management Authority and the agency, to be established, responsible for disaster response. 42 4.2.1.1 Restructuring to Increase the Coordination and Harmonization of Cash Transfer Programmes 162. In addition to the above overarching changes, the government is also committed to reviewing the current institutional structure in an attempt to increase coordination among and harmonization of the cash transfer programmes. The key objectives of this restructuring will include ensuring that there is sufficient capacity for programme implementation, enhancing cross-learning and coordination across programmes, and making use of corporate functions81 where appropriate. 163. Although no definite decision has yet been taken, it is likely that in future there will be two secretariats, one for the HSNP, which is likely to remain the responsibility of a separate ministry82, and one for the programmes currently managed by the MGCSD. In the initial stages of the NSNP, all of the programmes will continue to be implemented separately but with greater cross learning and more use of best practices. Over time, however, it is expected that they will begin to pool certain functions such as the contracting and management of payment agents, the management of grievance procedures, and possibly even targeting. 164. There is already evidence that cross-learning is taking place. The CT-OVC has incorporated lessons learned from the HSNP to improve its grievance procedures (section 3.5) and in its use of a bank as one of its payment service providers (section 3.4.2). The HSNP is now field testing a new targeting approach that incorporates a proxy means test adapted from the one used by the CT-OVC alongside its existing community-based targeting methodology (section 3.2.1.2). The OPCT and PWSD-CT are making much greater use of payment providers to deliver cash transfers following the example set by both the CT-OVC and the HSNP rather than making payments through the District Treasury. The new single registry and the development of the MISs are being formulated following a review of existing experience and with adaptations of the MIS system for the HSNP (section 3.3). 4.2.2 MINISTERIAL RESTRUCTURING 165. There is likely to be significant restructuring of the ministries after the forthcoming elections in order to comply with the new constitution, including a drastic reduction in their number to about half the current level. Both the MSDNKAL and the MGCSD are likely to be affected by this restructuring with responsibilities merged with or shared amongst other ministries. Any such changes clearly have implications for the NSNP, but the outcome of these changes will not be known until after the national election to be held, possibly in two rounds, beginning in early March 2013. 166. The CT-OVC and the HSNP have experienced similar shifts in their institutional arrangements in the past. The CT-OVC was managed by the Ministry of Home Affairs prior to 2009. Following restructuring, the team responsible for the programme’s day-to-day implementation and oversight was shifted almost in its entirety to a new ministry with only minor disruptions to the implementation of the programme. Similarly, the HSNP moved from the Ministry for Special Programmes in the Office of the President to the new MSDNKAL when it was formed in 2008. Therefore, while it is important for development partners to monitor the restructuring closely and to participate in relevant discussions, the changes should not be a serious cause for concern and may in fact provide opportunities for better coordination among the various components of the NSNP and between components and other relevant development interventions. 4.2.3 DEVOLUTION 167. The new Constitution of Kenya, passed in 2010, outlined a new decentralized structure of government in Kenya. Prior to this, the country was divided into eight provinces, each of which was 81 These may include contracting of payment service providers, grievance mechanisms, the single registry, some monitoring and evaluation functions, amongst others. 82 This separate ministry is currently the MSDNKAL, but following the ministerial restructuring may change. 43 composed of between three and 14 districts. 83 Since the introduction of the new constitution, the country is comprised of 47 counties with increased executive powers. Following the forthcoming election, each county will be led by a governor elected by the citizens of the county who will then have the authority to appoint the members of the county executive.84 168. As a consequence of the steps already taken in developing the new counties’ structures, a mixture of district-level staff (District Children’s Officers and District Gender and Social Development Officers) and county-level staff (County Coordinators) is now involved in implementing the cash transfer programmes (in particular the CT-OVC, the UFS-CT, the PWSD, and the OPCT). This is an evolving situation and there are therefore no nationwide standards governing country and district staffing composition. In addition to the evolving county structure, the number of districts continues to change. This state of flux is likely to continue until several months after the March 2013 election. 169. The increase in the executive power of counties is not expected to have any significant implications for the implementation of cash transfer programmes as social security remains a national-level function according to the new Constitution. However, there may be implications for district-level and county-level staff who may have dual accountability requirements, with some functions devolved to the county level (with reporting to the county executive) and others, such as cash transfers, remaining a national level function. Furthermore, county governors and members of the county executives are likely to take a considerable interest in the coverage and implementation of safety net interventions, which may also result in additional demands on district and county-level staff. 170. The full consequences of the devolution process, including new budgeting systems, have yet to be realized. There is likely to be some lack of clarity in the early months and years while the transition to the new system is completed and the necessary adjustments are made. 4.3 Incentives 171. The current Kenyan government has shown its commitment to social protection and cash transfers by adopting a new National Social Protection Policy and by the considerable and growing financial investment that it is already making in cash transfer programmes. As already indicated in section 2.2, the government has made the NSNP a high priority on its agenda. 172. There are several active advocates for social protection in the government and Parliament. The Prime Minister pushed for the development of the Urban Food Subsidy Cash Transfer, and many MPs have been supportive of the expansion of the different cash transfers and are demanding that their constituencies receive adequate coverage. However, the ministries within which the programmes are housed are not influential and are therefore not very effective in advocating for increased resources or timely release of funds. 173. The government’s system of performance contracts provides incentives for individuals working within the ministries to produce results. Performance contracts for each ministry are set at the beginning of the reporting period with targets agreed between the ministry and the Government of Kenya 85 and these contracts are reviewed at the end of the period. Performance contracts have a 83 In 1999 there were 69 districts. This number rapidly increased between 2007 and 2009, reaching 256 by June 2009. In 2009, the High Court ruled that there was a need to reduce districts to the original 47 outlined in the Districts and Provinces Act of 1992 (The Standard, September 4, 2009). Following the adoption of the new constitution in 2010, those 47 districts became the basis for the new counties, and the Office of the President was given a new mandate to create new districts as the President deems necessary. As a result, there are now more than 298 districts. 84 The Governor has the mandate to appoint the county executive, but members will also need to be approved by the county assembly. Executive members cannot be members of the county assembly. 85 Represented by the Permanent Secretary, Secretary to the Cabinet and the Head of the Public Service. 44 cascading approach with the targets agreed in the performance contracts of the overall Ministry cascading through the staffing structure. The current performance contract for the MGCSD includes indicators related to the number of beneficiaries covered by each cash transfer programme and its expansion. Ministries are ranked according to their performance, and these results are published and reported on by the media.86 174. There are numerous opportunities for communities and beneficiaries to be involved in decision-making within the HSNP and CT-OVC. These opportunities include community targeting (in the HSNP), the community verification of beneficiaries, and membership of Rights or Beneficiary Welfare Committees. When combined with an improved grievance system, this community and beneficiary involvement is designed to ensure that service providers can be held to account. These opportunities are weaker in the remaining three programmes and need to be strengthened, as does communication about the objectives and core procedures of each of the five programmes. 4.4 Assessment of Adequacy 175. The establishment of a National Social Protection Council and support to it through the Social Protection Secretariat will raise the profile of the social protection sector and enhance the coordination and oversight of activities. It is vital that the efforts that are already underway to strengthen the Social Protection Secretariat are continued and enhanced. It is also important that the relationship between the secretariat and the agencies responsible for the day-to-day management of cash transfer programmes are strengthened and formalized. 176. The performance of HSNP and CT-OVC in their previous phases indicated that their institutional structures are functioning well and that lines of accountability are working. With regard to the HSNP, much of the functional elements of the programme are sub-contracted to CSOs and the private sector with only limited oversight by the government. Therefore, there is a need to enhance the government’s ownership and oversight of the HSNP programme. For example, although the HSNP secretariat is located within the MSDNKAL, it is entirely staffed by contract staff and there are no civil service counterparts; and there is no government financing of the programme. Building ownership, introducing government financing, bringing the programme “on-budget” (see section 5.1) and ensuring a staffing structure within the government will be critical for ensuring the programme’s sustainability. 177. There is a need for systems and structures to be put in place for the remaining three programmes. A common cash transfer secretariat, if implemented, would facilitate common oversight mechanisms and would constitute a national-level operational unit for the programmes managed by the MGCSD. However, it will be critical to ensure that this secretariat is adequately staffed and that appropriate systems and manuals to guide them are developed. A likely consequence of such an approach would be a dedicated staff member at country and or district level to work on cash transfer programmes. Such an approach would resolve the current challenge faced by district staff of competing priorities between cash transfer responsibilities (often seen as an add on to core work responsibilities) and their other tasks. 178. Assessments of the various cash transfer programmes consistently refer to capacity gaps at the district and community level, in particular the lack of time and staffing allocated to support targeting (see section 3.2.1.2).87 The assessments also highlight the lack of financing and equipment. The external monitoring for the CT-OVC has highlighted the lack of a budget for programme activities at either the district or the community level,88 while an assessment of the MGCSD’s social protection capacity found a lack of equipment and, more seriously, vehicles at the district level.89 86 Office of the Prime Minister (2012) and Government of Kenya (2012) 87 Calder et al (2011)OPM (2012), and Concern (2012) 88 Kimetrica (2010b) 89 MGCSD (2010) 45 179. Chiefs and sub-chiefs often play critical roles in programme implementation particularly in relation to targeting and communication flow, but their responsibilities are not formally recognised in programme documentation. There is concern that chiefs and sub-chiefs can have excessive influence over key programme decisions but the lack of guidance concerning their role exacerbates this. In the future, their role should be recognised and documented with appropriate checks to mitigate any undue influence (such as better awareness of rights and responsibilities as well as mechanisms for complaints and appeals). 180. All programmes rely heavily on voluntary efforts by community members. The time demands on these unpaid volunteers are significant, and there are frequent reports of them asking to be reimbursed in some way, for example, in the form of stipends or credit for their mobile phones.90 This is a challenge for the sustainability of relying so heavily on volunteers and increases the risk that these volunteers will demand payments from beneficiaries or potential beneficiaries. 181. A full functional review of the NSNP will be conducted in mid-2013. This review should assess both national and local-level capacity requirements and constraints and reflect the ongoing shifts in district and county level management (as well as the expected restructuring at the ministerial level). This review should carefully consider the existing capacity requirements of the NSNP and the increasing demands that will result from scaling up the constituent programmes and from enhancing systems and should make its recommendations accordingly. It is expected that the recommendations will touch on many of the points mentioned above but will also consider:  The overall programme organogram and how coordination will be improved among the programmes managed by the MGCSD and between these four programmes, the HSNP and the component focused on scaling up the NSNP in response to shocks.  The need to ensure adequate staffing for the programme including the Social Protection Secretariat, the various departments at the national level responsible for programme implementation, and, most importantly, at the county and district levels. The review should make specific recommendations on staffing levels and on the required skill sets of these staff.  Any need for short-term or medium-term technical assistance to support the functioning of the implementation units.  The consequences of devolution for programme implementation and the potential for increased competing demands on district and county-level staff. This will include what are the roles and responsibilities of the county offices, what will happen to the current district structure and how does reporting work between levels. 4.5 Management of Error, Fraud, and Corruption 182. Every cash transfer programme aims to provide the right amount of benefit to the right recipient on time. However, the evidence shows that some level of error, fraud, or corruption is inevitable in all cash transfer programmes. Figure 9 below illustrates the typology of error, fraud, and corruption (EFC) in a typical benefits system. Figure 9: Typology of Fraud and Error Source: Van Stolk and Tesliuc (2010) 183. Examples of the nature and type of EFC risks faced by cash transfer programmes through different stages of their lifecycles are described in Table 12 below. 90 Kimetrica (2012a) and Calder et al (2011) 46 Table 12: EFC Risks at Different Stages of Programme Lifecycle STAGE RISK TYPE Registration/ i. Beneficiary disguises the household’s economic position Fraud Enrolment ii. Collusion by local representative to facilitate a Corruption household’s acceptance into the programme iii. A household is wrongly enrolled despite its declared Official Error circumstances iv. A household is enrolled in multiple programmes Fraud or Corruption/or possible Official Error v. Ghost beneficiaries are enrolled Corruption vi. A household gives a genuine, but mistaken, perception No fault of being in extreme poverty Payment i. Beneficiary is short-changed by the PSP Corruption ii. Beneficiary is short-changed by the caregiver Corruption iii. The agent or programme manager demands a cut of the Corruption transfer iv. The wrong amount is paid Official Error Change i. Undeclared change – the household was not aware of the Beneficiary Error need to report the new information ii. Undeclared change – the household deliberately withheld Fraud the new information iii. Local representative was told but took no action because Official Error unaware of materiality iv. Local representative told but took no action although Corruption aware material v. Failure to act when maximum age for programme Official error reached Exit i. Exit trigger was mistakenly not reported by the Beneficiary Error beneficiary/recipient ii. Exit trigger was deliberately not reported by the Fraud beneficiary/recipient iii. Exit trigger was deliberately not reported by local Corruption representative Replacement I. Incorrect erroneous selection of replacement household No fault from waiting list ii. Deliberate incorrect selection of replacement household Corruption from waiting list Source: Barr (2012) 184. Evidence generated by KENAO audits and programme evaluations suggest that rates of EFC are low in the programmes being assessed, but there have been some incidents (often reported anecdotally). These include beneficiary households who do not meet eligibility criteria (for example, they do not meet the age requirement of the OPCT or are not taking care of an OVC despite benefitting from the CT-OVC); allegations of discrimination, corruption, and nepotism; suggestions of ghost beneficiaries in the UFS-CT; some beneficiaries participating in more than one cash transfer programme; and changes in the composition of beneficiary households that are not reflected on the payroll. Evidence of EFC in the HSNP is low, but this programme’s operational manuals are more detailed than those for the other programmes and include protocols for case management including the reporting of changes in beneficiary circumstances.91 91 Barr (2012) 47 185. A number of existing and planned controls are in place in order to minimize the opportunities for error, fraud, and corruption in the various stages of programme life-cycle described in Table 12 above. These are outlined, by stage of programme implementation, in Table 13 below. Table 13: Current and Planned Controls to Minimize Error, Fraud, and Corruption Stage CURRENT PLANNED Registration/ - Targeting methodology - Stronger ID verification Enrolment - Community involvement/knowledge - The MIS will reject information is potentially a strong preventive inconsistent with CT eligibility (e.g. a child measure, but local practice is over 18 in the OVC application). unreliable. - Single registry database of all beneficiaries - There is a question on the application will enable duplicates to be identified. form about enrolment in other CT - The single registry will enable IDs to be programmes. cross-checked against the IPRS database - Corruption can be reported on the - Improved and standardized complaints and MGCSD website but procedures are grievance channels vague. Payment - Having the same flat rate for all CTs - The MIS will flag irregular amounts helps the beneficiary/recipient to know outside norm to be scrutinized before how much to expect and reduces scope payment. for error and short changing by - Extension of electronic payments agents/carers. - Extension of two-factor ID authentication - Pre-payment payroll checks at the for collection of payments national level - Post-payroll checks - Post-payment reconciliations - Improved and standardized complaints and - Termination after three successive grievance channels failures to collect transfers - Internal audit scrutiny - Complaints and corruption reporting Change/Exit Form for notifying change exists - The new MIS will have automatic triggers Beneficiaries and local committee to cease making payments such as death or members told about notifying change the maximum eligible age has been reached. of circumstances - Retargeting/recertification exercises Replacement - Waiting lists controlled at the district level - Revalidation of households before enrolment 186. These current and planned actions represent significant controls against fraud and error, but further additional actions are recommended. These include:  Given the inevitability of some degree of EFC occurring in cash transfer programmes, it is important that there is recognition and ownership of these risks at the ministry level and clear allocation of responsibility for controlling EFC at each tier of administration.  Operational manuals and procedures should be updated to reflect EFC risks and controls and be communicated to local officials to raise their awareness of EFC.  There is also a need to explain what actions constitute EFC to beneficiaries and caregivers and to increase their understanding of programme systems and procedures.  There should be more analysis of complaints about specific aspects of programme delivery in order to identify and address any systemic problems and any potential cases of fraud and corruption.  There needs to be a better system for investigating possible cases of fraud and corruption, particularly by collecting key details through the complaints process. 48  The management of waiting lists needs to be improved, including ensuring that an audit trail exists on how replacement beneficiaries are selected and carrying out spot-checks of the accuracy and appropriateness of the process.92 5 BUDGET PROCESS AND EXPENDITURE FRAMEWORK 5.1 Budget Process and Structure 187. How government and development partners’ resources are allocated is guided by the government’s long-medium and short-term priorities. These priorities are documented in the Vision 2030 document, the medium-term expenditure framework (MTEF), and the annual budget estimates. Vision 2030 lays out the government’s long-term priorities (covering the period from 2008 to 2030), while the MTEF provides a more medium-term plan for the allocation of resources among different sectoral priorities. The formulation of the Annual Budget Estimate is guided by the MTEF and a budget policy statement. This process is described in more detail below. 188. There are four key stages to the budget preparation process:  The Treasury issues a budget circular that lays out the government’s medium-term fiscal strategy, sectoral resource ceilings, and other information such as the budget preparation calendar and the composition of sector working groups.  Sector working groups prepare their sector budget proposals guided by the budget circular, ministerial expenditure and sector budget reviews, the Vision 2030 Medium-term Plan, and relevant sector policies and strategies.  The Macro Working Group supported by the Economic Affairs Department in the Ministry of Finance prepares a budget policy statement. This is a three-year rolling93 document outlining the key macroeconomic information and macro decisions needed to guide the budget development process. It reports on the planned and approved expenditures executed in the prior two years by each major sector budget line, presents a baseline and ceiling budget figure for the forthcoming year to guide the sector working groups in their work, and makes provisional budget projections for two years ahead. This budget policy statement must be approved by Parliament.  Each sector then prepares a detailed budget for their ministries, departments, and agencies guided by the budget policy statement. These detailed budgets must also be presented to and passed by Parliament before the final budget is completed. Figure 10: Timeline and Key Activities for Budget Preparation Source: Ndung’u (2012) 92 For further details, see Barr (2012) 93 The 2011 document provided budget information for the years 2011,2012 and 2013, while the 2012 document provides information for the years 2012,2013 and 2014. 49 189. The budgets for the cash transfer programmes are prepared by the sector working group responsible for social protection, culture, and recreation. This group is responsible for preparing budgets for the following ministries:  Ministry of Heritage and Culture  Ministry of Gender, Children, and Social Development  Ministry of State for Special Programmes  Ministry of Youth Affairs and Sports  Ministry of State for Development of Northern Kenya and Other Arid Areas 190. Within the sector budget proposals, budgets are classified according to each of the above ministries and by whether the planned expenditure is for recurrent costs or development. The development budget heading is further divided as to whether it is capital expenditure or other development expenditure. Development projects and capital investments are budgeted under the development budget heading, whereas operations and maintenance costs are included under the recurrent expenditure budget. Funding for programmes financed by development partners is budgeted under the development budget heading regardless of whether it is for a capital investment or for operations and maintenance. 191. The budget proposals for the cash transfer programmes are based on the estimated number of beneficiaries to be covered or enrolled and, in theory, on the estimated operational costs and monitoring resources needed to support this caseload. In practice this only happens for the CT-OVC, non-transfer expenses for the other three programmes are drawn from the Ministry’s general operation expenses (as discussed further below). The budget is split between the development heading and the recurrent heading. While all donor financing for cash transfers is categorised under the development budget, as described above, government financing is split between the two headings. Although cash transfers would typically be categorized as recurrent expenditure (and the majority of cash transfers are classified as such), the MGCSD has lobbied for and gained additional resources from under the development heading. 192. Within the ministry budgets, the projected expenditures are then broken down further according to a chart of accounts. The chart of accounts in all ministries and programmes follows a standard government format that includes a large range of line items that cover, for example, staffing, communications, and travel. There is a programme-specific chart of accounts for the CT-OVC, which makes it possible for expenditure to be easily monitored by line item separate from other expenses incurred by the MGCSD. The other programmes managed by the Department of Gender have no programme-specific chart of accounts but instead are subsumed under the main MGCSD chart of accounts. While it is still possible to track actual expenditure on transfers to beneficiaries, this makes it impossible to track other expenses related to the implementation of these programmes as they are merged with the rest of the MGCSD budget. 193. The way in which the budget process is structured means that it is not possible to guarantee a predetermined multi-year allocation to either the social protection sector or to a specific programme within the sector. All projected funding in the MTEF is subject to revisions if macroeconomic conditions in Kenya change. It may be possible to tie the P4R financing directly to the sector and to specific cash transfer programmes, but this would mean that any budget provided by the government in anticipation of financing from the World Bank would be over and above the MTEF allocation and therefore could not be sustained beyond the time when World Bank financing is available. Such an approach could therefore undermine the objectives of the programme, one of which is to promote an increase in the government budget allocation for cash transfer programmes. 194. There are several reasons to suppose that the funding for the NSNP will be sustainable over the long term. First, the World Bank or other development partners can participate in discussions of the sector working group. Second, the desire to meet the disbursement-linked indicators (DLIs) is likely to prompt some use of the discretionary powers of the Budget Department and the Treasury 50 where necessary. Third, the government has a genuine and proven commitment to scaling up social protection and to funding the programme’s expansion. It will be important for the World Bank to ensure that the incentives provided by the DLIs are maintained in the later years of the programme by scheduling planned disbursements appropriately. 5.2 Programme Financial Sustainability 195. Cash transfer programmes in Kenya currently cover an estimated 1.6 million beneficiaries in 279,843 households throughout the country. In terms of financing, the current fiscal year allocation to these programmes amounts to Ksh 8.04 billion, of which the government is providing 39 percent (Ksh 3.15 billion) and development partners are providing 61 percent (Ksh 4.89 billion). 196. The Government of Kenya has stated its ambition to increase the coverage of cash transfer programmes to provide support to an increased proportion of households with members who belong to specific vulnerable groups. Achieving this goal will require the government to gradually increase its financing for cash transfer programmes. 197. There are indications that the government will not only continue but will increase the current level of public financing for cash transfers given its strong political commitment to these programmes. Importantly, however, there is a need to balance the aim of increasing coverage to progressively realize the right to safety net support with the need to ensure the long-term sustainability of the NSNP. As a result, the government proposes to first move all of the funding for cash transfers from whatever source “on-budget” and then to progressively take over development partner funding (including funding of the HSNP), while simultaneously increasing the number of people who benefit from these programmes. As this will take some time, additional resources will still be required from development partners in the short to medium term to support the proposed expansion of the NSNP. Figure 11 below indicates the government’s plans for expanding the number of NSNP beneficiary households while increasing domestic financing.94 Figure 11: Number of Beneficiaries per Programme Year by Funding Source 198. As Table 14 below shows, this increasing financing remains within achievable levels when expressed as a percentage of GDP and is a small proportion of the anticipated increases in government revenue expected as a consequence of GDP growth. 94 The figures presented in Figure 11 and Table 14 are drawn from the current version of the programme Medium Term Expenditure and Financing Framework (Version 9 April 2nd). This framework will continue to be reviewed and revised over the coming months. 51 Table 14: Affordability of Government Financing of the NSNP FY 12/13 FY 13/14 FY 14/15 FY 15/16 FY 16/17 GOK Financing 3,157,000,000 4,322,500,000 6,743,374,200 9,405,869,580 13,665,673,024 DP Financing 4,886,659,087 6,215,209,087 5,375,209,087 4,906,084,087 3,658,211,588 Total Financing 8,043,659,087 10,537,709,087 12,118,583,287 14,311,953,667 17,323,884,612 GOK financing as % 0.10% 0.10% 0.14% 0.18% 0.25% of GDP 5.3 Funding Predictability Table 15: Expenditure Review for the MGCSD (KSH Millions) Approved Estimates Actual Expenditure 2008/09 2009/10 2010/11 2008/09 2009/10 2010/11 Gender and Social Development 1,051.2 1,086.3 1,988.2 1,051.2 1,086.3 1,759.4 Children’s Services 2,010.1 591.8 3,568.2 2,010.1 591.8 4,118.3 General Administrative Services 0.0 620.2 578.3 0.0 620.2 606.9 TOTAL 3,061.3 2,298.3 6,134.7 3,061.3 2,298.3 6,484.6 Source: Ministry of Finance (2012) 199. As Table 15 above shows, overall MGCSD95 spending has largely been in line with budget projections. However, disbursements of funds from the Treasury throughout the year do not always meet the cash flow needs of the four cash transfer programmes managed by MGCSD. There tend to be two main challenges. First, the closing of the accounts at the end of the year frequently results in a delay in the first disbursement of the new financial year. Second, short-term cash flow constraints within the exchequer, combined with prioritization of other expenditures, can result in delays in disbursements to particular ministries. As mentioned in section 3.4.3 above, this can in turn delay the delivery of payments to beneficiaries and can also affect other programme activities such as targeting. 200. Salaries and emoluments, debt servicing, and public servant pensions are protected from the delays caused by the closing of the financial year and are prioritized when the exchequer is experiencing cash flow constraints. This is because these expenditures are paid directly from the consolidated account prior to releasing the allocations to different ministries on the basis of appropriations. 96 As a consequence, these expenditures are honoured prior to any other prioritization process which may come into effect in the event of cash flow limitations. Therefore, policymakers should explore the option of reclassifying cash transfers as salaries and emoluments as a way to overcome this constraint. 5.4 Adherence of the Budgeted Programme Expenditure and Execution to Government Priorities 201. As stated above, overall MGCSD97 spending is largely in line with agreed budgets, which indicates that the ministry’s planning and budgeting is largely realistic and has enough capacity to spend according to plans. However, as Table 16 below shows, there is a tendency for the cash transfer programmes to underspend their budgets. 95 To date HSNP has been entirely donor financed, with predictable financing provided directly to sub- contracted implementing partners. 96 National Council for Law Reporting (2010) 97 To date the HSNP has been entirely donor financed, with predictable financing provided directly to implementing partners. 52 Table 16: Expenditure against Budget 2011/12 for Programmes Managed by the MGCSD in millions KSH CT-OVC OPCT, PWSD-CT & UFS-CT * Budget Expenditure % Spent Budget Expenditure % Spent Cash Transfer 3,293.86 2,862.39 87% 1,583.00 1,434.05 91% Transfer costs 79.19 48.77 62% Consultancy 134.29 50.65 38% Operating costs 397.71 287.43 72% 3,905.05 3,249.25 83% 1,583.00 1,434.05 91% Source: Finance information provided by Department of Gender and Social Development and Department of Children’s Services of the Ministry of Gender, Children, and Social Development. Note: * Information on overhead costs is not accessible due to the current chart of accounts. 202. Although the percentage of underspending may be higher on non-transfer expenditure, the major budget component is the cash transfer budget and so this contributes most to the overall underspending. The main reason for this underspending is delays in targeting which leads to knock on delays in cash transfer payments. Thus underspending is higher in those programmes that expanded their coverage during 2011/12. 203. While underspending is not excessive, there is a need to improve the planning and execution of expenditure, particularly with regard to ensuring that operational systems are in place for the smooth and efficient flow of transfers to targeted beneficiaries. This issue is explored further in section 5.5 below. Furthermore, there is an urgent need to adjust the chart of accounts for the MGCSD to ensure that appropriate budget lines are available to plan and track operating expenditures. This will also need to be put in place for the HSNP once the programme moves on- budget.   5.5 Efficiency of Programme Expenditures 204. Figure 12 below indicates the cost efficiency of the CT-OVC over the past five years (data are not currently available for the HSNP and analysis is not possible for the remaining three programmes due to the chart of accounts not allowing tracking of overhead costs). As Figure 12 shows, costs have declined over time as systems are put in place and efficiencies gained. Figure 12: Cost-efficiency of the CT-OVC 205. By 2012 the cost to transfer ratio for the CT-OVC and HSNP had declined to 1.14 and 1.2198 respectively; with cost predicted to fall further during the NSNP to an average of 1.15 by 2018. As the table below demonstrates, this is within the normal parameters of cash transfer programmes: 98 Informal communication from DFID. 53 Table 17: Cost to Transfer Ratio for Selected Programmes Programme CTR Source Progressa, Mexico, 1997-2000 1.058 by 4th year Caldes et al, 2004 LEAP Ghana 2007-10 2.11 by 4th year White, 2011 Dedzo Safety Nets Pilot, Malawi 1.65 White & McCord, 2006 rd Transfers Pilot, Zambia 1.09 by 3 year White & McCord, 2006 PSNP, Ethiopia 2010-11 1.25 * White and Ellis, 2012 Note: All sources quoted in White and Ellis (2012) * This excludes the costs of items procured to support the implementation of public works, but includes the cost of transport and handling of food transfers, which make up a proportion of the programme. 206. The cost efficiency of social assistance programmes is dependent on a number of factors:  The complexity of the programme, in particular the inclusion of co-responsibilities or public works.  The size and spread of a programme and the impact this has on the ability to take advantage of economies of scale (the administrative cost per unit is likely to be higher for a programme supporting 23 beneficiaries per location than one supporting 150).  The cost of key administrative components of the programme such as targeting and monitoring systems or the payment mechanism.  Whether or not the programme is in a start-up phase requiring investment in systems and procedures. 207. The lack of programme specific charts of accounts for the OPCT, PWSD, and UFS-CT means it is not possible to identify administrative costs associated with these programme, but are known to be minimal given the limited budget provision to districts, the low spending on technical assistance and the limited government financed monitoring and evaluation. These low overhead costs are a reflection of the limited investment that there has been in setting up effective management structures, providing appropriate oversight and ensuring that those implementing the programme have the required knowledge and skills. As a consequence, the overhead costs of the CT-OVC and the HSNP programmes are expected to continue to decline while the overhead costs for the OPCT, the PWSD- CT and the UFS-CT should increase substantially towards the start of the NSNP, as the necessary investments are made before flattening out or even decreasing. 6 RESULTS FRAMEWORK 6.1 Overview of Results Framework 6.1.1 EXISTING RESULTS FRAMEWORKS 208. The results framework for the NSNP should be consistent with results frameworks of the individual programmes. However, only the HSNP, CT-OVC, PWSD-CT and OPCT have draft results or monitoring and evaluation frameworks available. The UFS-CT has draft programme objectives set out in a draft programme document, but no logical or monitoring and evaluation framework. 209. The HSNP phase 1 is ending and HSNP+ (or phase 2) will begin in 2013. There was no logical framework or detailed monitoring and evaluation framework for phase 1, though there were extensive monitoring and evaluation activities (see below). The HSNP+ (phase 2) draft monitoring and evaluation framework99 notes that there is still no logical framework available but does specify primary and secondary expected impacts and, for most impact areas, instruments to provide data from 2012 to 2018. Primary expected impacts include increased consumption expenditure, increased food security, reduced malnutrition, and increased asset retention and accumulation. Secondary expected impacts include indicators around health, education, livelihood diversity, vulnerability to shocks, 99 Pinney (2012) 54 financial savings, price stability, and well-being of vulnerable groups. The framework does not specify inputs, activities or outputs. Not all of the impact areas have clear indicators attached to them. 210. The CT-OVC draft logical framework100 includes the programme objective (‘provide a social protection system to Orphans and Vulnerable Children), purpose (‘Strengthen capacity of families to care and protect the OVC, and to promote their human capital development’), outputs (‘Beneficiary households (poor households living with orphans and vulnerable children) reach an income/expenditure level that exceeds the national ultra poverty line through Provision of regular and predictable’ [transfers]) and activities. Under each component an ‘intervention logic’ specifies an objective (e.g. increase household food consumption and promote nutritional status), for which there is a verifiable indicator, a means of verification, and basic assumptions. This framework is reasonably comprehensive, but is still in draft, does not specify inputs or link activities to outputs, and contains some inadequately specified indicators. The CT-OVC programme manual sets out expected monitoring and evaluation activities but does not refer to the logical framework. 211. The OPCT and PWSD-CT share a draft results framework, set out in an operational review of both programmes.101 This draft framework is comprehensive and includes an overall goal (Reduce poverty and vulnerability in Kenya by providing support to the elderly/persons with severe disabilities), impacts (improved i) welfare, and ii) resilience of beneficiary households, and iii) increased economic and social effects), outcomes (e.g. increased food security), an intermediary objective (provision of a sustainable, equitable, national SP programme), outputs and activities within each output. Each goal, impact, outcome, output and activity has a clearly defined indicator, disaggregation, means of verification, periodicity, assumptions and targets. However, this remains in draft, has not yet received comments from the government and still contains gaps. It is expected that these will be completed before the NSNP P4R starts. 6.1.2 NSNP RESULTS FRAMEWORK 212. A draft NSNP results framework has been developed in close collaboration between the government and World Bank task team, based on the existing and draft programme results frameworks. It was presented at a stakeholder workshop and revised following a round of comments from government and development partner staff. The current NSNP results framework remains in draft pending a further round of comments and revisions. 213. The draft NSNP framework (included as Annex 1) links inputs, activities, outputs, outcomes and impact. The full draft specifies indicators, baselines and targets, frequency, means of verification (data sources) and responsibility. The draft contains 39 indicators. These indicators are largely Specific, Measureable, Attainable, Relevant and Timely (SMART). However, some indicators lack baselines and targets are still to be agreed. Procedures for routine monitoring of the indicators are clearly set out and largely feasible, but further capacity needs to be built to ensure this (see below). 214. The overall goal or impact of the NSNP is ‘Reduced Vulnerability and Poverty in Kenya’, measured through changes in the hardcore household poverty rate estimated by the Kenya Integrated Household Budget Survey (KIHBS, which takes place every five years with the next iteration in 2013). Changes in the poverty rate cannot be fully attributed to the NSNP and the KIHBS cannot provide estimations of impact, but this is nevertheless important in the results framework since the NSNP should make significant contributions towards poverty reduction. More directly, the NSNP is expected to have outcomes in five areas within improved beneficiary welfare and two areas within improved beneficiary resilience: 100 ‘CT-OVC Programme Logical Frame Work’, no date. 101 Kimetrica (2013) 55  Output 1: Improved Beneficiary Wellbeing: Net change in beneficiaries’ i) consumption expenditure, ii) dietary diversity, iii) households’ members under 18 attending school, iv) morbidity, and v) self-esteem score.  Output 1: Improved Beneficiary Resilience: Net change in beneficiaries’ i) asset ownership index, and ii) reported engagement in income generating activities. 215. Based on previous evaluations of cash transfer programmes in Kenya, it is realistic to expect the NSNP to affect these indicators. However, impact and outcomes are not fully within the control of the programme, and may be affected by weather, inflation, changes to service provision, beneficiary behaviour and so on. Programme officers are aware of these risks, which are set out in the full monitoring and evaluation plan. Changes in outcome indicators need to be attributed where possible to programme activities, and will therefore be measured through impact evaluations of the programmes. Currently an evaluation of the CT-OVC is ongoing with a baseline in 2012 and reports on these indicators potentially available in 2014. DFID expect to finance an evaluation of the HSNP beginning in 2013. There are no other plans currently for rigorous impact evaluations of the programmes on which the results framework could draw, but these may be funded by development partners. In the absence of further impact evaluations, it is considered reasonable to use impact evaluations of the HSNP and CT-OVC as proxies for the impact of the NSNP as a whole, given the similarities of the programmes. 216. There are six output areas. These outputs are all seen as necessary for the achievement of the programme development objectives, but some capacity needs to be built among programme officers and decision-makers to demand and use monitoring information on these outputs (see below). The outputs are i) The most vulnerable and poorest households are enrolled in the NSNP; ii) NSNP beneficiaries receive appropriate, reliable and accessible payments; iii) Citizens are able to appeal and complain to improve programme performance; iv) A monitoring and learning system functioning, v) The cash transfer sector is harmonised and government managed, and iv) Programmes are responsive to shock. Each area has outputs and in some cases activity indicators. Outputs and activities are fully within the control of the government; in most cases they are within the control of the NSNP but in some cases (such as timely payments or the agreement of harmonisation plans) they also rely on other agencies of government such as the Treasury or the Cabinet. These will be tracked using the Management Information System (MIS) and single registry; a Programme Implementation and Beneficiary Satisfaction (PIBS) Survey conducted by an independent firm or firms, and spot checks from programme officers. 217. There are two input areas: i) Resourcing is adequate, affordable and sustainable; and ii) Staff and organisational capacity is adequate for efficient and effective implementation. These are also within the control of the programme and will be tracked through programme budget data, payrolls, and reviews. 218. The disbursement linked indicators fit within this system. Given that the NSNP’s results framework is highly detailed, the PfR will adopt a subset of these results (using 13 indicators), and this subset is given in Annex 2 of the PAD. 6.2 Monitoring and Evaluation System 6.2.1 EXISTING PROGRAMME ARRANGEMENTS 219. The institutional set up for the NSNP monitoring and evaluation (M&E) system is currently split across the five programmes, each of which has its own M&E arrangements. M&E arrangements in the four MGCSD programmes are broadly similar in practice, although the CT-OVC has better established M&E procedures and staffing than the other three. The HSNP M&E arrangements differ slightly from those of the MGCSD programmes. Across all the programmes, there are significant gaps in capacity for M&E, and limited public disclosure of monitoring information. 56 220. At the level of the new NSNP M&E institutional arrangements are currently quite new and are in development. The Social Protection Secretariat currently has one M&E officer who is a permanent civil servant, seconded from the Ministry of Planning but reporting to the head of the Secretariat. There is also currently an M&E specialist funded by UNICEF and reporting to UNICEF and the head of the Secretariat. Both staff have responsibilities across the social protection sector, not just for the NSNP. There is an M&E budget within the Secretariat budget for workshops, printing, and some limited travel. Currently demand for M&E information in the Social Protection Secretariat is quite weak as the Secretariat is dealing with rapid change across the sector and new institutional arrangements are being developed. There are limited demands on the Secretariat itself to provide information to the government or other partners. In addition, there should be monthly M&E reports for the Secretariat containing the monitoring workplan and activities achieved by the Secretariat M&E, but it is not clear that these are always submitted or demanded. 221. The CT-OVC has more developed monitoring and evaluation arrangements than the three DOGSD programmes, and these are set out in an operational manual. There is a dedicated M&E unit with four staff, which is adequate in terms of numbers and currently manages and supervises independent organisations conducting evaluations and monitoring surveys. However, the unit may require further capacity development to ensure that M&E activities are undertaken efficiently and to quality standards. The CT-OVC operational manual specifies three aspects of the M&E system: i) Internal monitoring to track inputs and activities, aiming to identify problems and allow the introduction of corrective measures, and producing regular monitoring reports to be shared with development partners; ii) External monitoring including a spot check, citizen report card survey, and independent grievance mechanism to a) ensure that the operational manual is followed, b) monitor risks and ensure that they are being mitigated, and c) assess beneficiary and public satisfaction; and iii) An operational and impact evaluation providing information on Institutional relationships and coordination, targeting effectiveness, operational processes, beneficiary and others’ perceptions of the programme, cost and cost-effectiveness, and unanticipated impacts. 222. These specifications are largely followed. The programme MIS produces bi-monthly reports on key payment-related indicators. The Technical Working Group (TWG) has established an M&E Task Force which meets every week to review the progress of CT-OVC M&E activities, and has so far focused on complaints and grievances, beneficiary welfare committees, the citizen service charter and the co-responsibilities pilot. Programme officers (including but not limited to the M&E unit) conduct field visits to 14 districts per payment cycle to follow up on specific issues. Each visit has its own TOR and reports are shared only with the OVC Secretariat and the Permanent Secretary but not more widely. The contractors of the external grievance system provide yearly summary statistics on complaints. There are regular audits and reviews under the Joint Review and Implementation Support (JRIS) mission, conducted by development partners and the GOK every six months, with the latest conducted in November 2012 and producing an aide memoire. Regular external monitoring is conducted, including i) Spot checks following a guidance document prepared by Kimetrica 102 , conducted each quarter to ensure compliance with operational manuals with three reports submitted103, ii) A community level monitoring system through focus groups with beneficiaries, non- beneficiaries and traders to assess programme knowledge and unintended market and social impacts of the programme.104 This was discontinued and replaced by beneficiary household surveys in the consolidated system, iii) Consolidated external monitoring and spot checks replacing the separate system, with a first draft report submitted in October 2012.105 Several external impact and targeting assessments of the CT-OVC have also been conducted, including: i) An evaluation of impact and targeting by Oxford Policy Management (OPM), with the baseline and targeting assessment submitted 102 Kimetrica, (2010a) 103 Kimetrica, (2010b), Kimetrica, (2012d), Kimetrica, (2012c) 104 Kimetrica (2010b), 105 Promin Consultants et al. (2012). 57 in 2008 and the final impact and operations report submitted in 2010;106 ii) A review of targeting by Development Pathways, submitted in 2011;107 iii) An impact assessment by the University of North Carolina and the Transfer Project, with various draft and final studies prepared on specific subjects, but no single evaluation report;108 and iv) an ongoing evaluation conducted by Mott Macdonald, still in baseline data analysis stage. 223. The CT-OVC has abundant M&E information. The quality of the external monitoring and impact evaluations is high, but these are costly and are not always used by programme officials. The quality of the information in the MIS and from the internal monitoring is less certain, and the provision of information and reports from these sources is irregular. The arrangements could be strengthened in at least four areas. These areas are common to all MGCSD NSNP programmes. First, the system could be improved by conducting regular and more systematised internal monitoring. The external monitoring report noted that DCOs mostly did not comply with the operational manual’s procedures on monitoring. There could be greater regular monitoring of outputs (rather than inputs) using the MIS and analytical work relating to understanding why errors occur. The draft aide memoire from the November 2012 noted that the OVC Secretariat was not yet producing regular bi- monthly implementation progress reports that included both data from the MIS and narrative reporting. This was put down to lack of capacity. M&E officer field visits could follow a clearer format and reporting schedule. Second, the MIS could be strengthened both by moving to an electronic system (this is currently happening with support from development partners) and by building capacity at the sub-county level to feed quality monitoring information into the MIS (which requires additional person days, transport resources, and computers at sub-county level). Third, demand for M&E information could be strengthened, as in all the programmes. Fourth, the external monitoring and impact evaluations could be more focused on the needs and demands of programme managers. 224. The OPCT, PWSD-CT and UFS-CT have similar reporting arrangements operating across the Department of Gender and Social Development (DOGSD). There are no dedicated monitoring and evaluation officers for any of these programmes. Rather, the seven programme officers in the Department share monitoring and evaluation responsibilities amongst their other duties. This may limit the attention devoted to M&E by these programmes. Dedicated M&E officers for each programme would improve the regularity and quality of M&E. District (soon sub-county) Gender and Social Development Officers (DGSDOs) and County Coordinators also participate in monitoring and evaluation, and the programmes have also had external assessments. The OPCT and PWSD-CT have operational manuals that set out more detailed M&E procedures, while the UFS-CT does not have a manual. All three programmes have had recent reviews of operations and impact (though not rigorous impact evaluations). These reviews contain much useful operational information, but demand for this information from programme officers is limited. 225. The current sources of monitoring and evaluation for the UFS-CT, PWSD-CT and OPCT are: i) Reports from DGSDOs before and after each payment cycle sent by email or post, with no fixed format; ii) Periodic (at most quarterly) reports from County Coordinators and/or DGSDOs; iii) Ad hoc communication from DGSDOs; iv) Ad hoc interrogation of programme MIS by programme officers; v) Ad hoc field visits by programme officers, and vi) External assessments, specifically a) an implementation evaluation by Concern, covering key impact areas and operations,109 b) a review of the OPCT and the PWSD-CT, carried out by Kimetrica,110 and c) an impact assessment of the OPCT by Agency for Development Research Across all three programmes, there is very limited aggregation of monitoring indicators. 106 OPM (2010) 107 Calder et al. (2011) 108 Taylor et al. (2012), Solomon Asfaw et al. (2013), Kenya CT-OVC Evaluation Team (2012a) and Kenya CT- OVC Evaluation Team (2012) 109 Concern (2012) 110 Kimetrica (2013) 58 226. The UFS-CT has the weakest M&E arrangements of these programmes, reflecting that it is the smallest and newest programme. The absence of any M&E procedures or data collection and reporting formats and schedules is a severe constraint to the quality of M&E information for this programme. The expansion of an electronic MIS to the UFS-CT will also improve M&E arrangements, but the capacity of district officials to provide data for the MIS will remain a key constraint unless strengthened. In addition, there is limited demand from programme officers for M&E information. 227. The PWSD-CT has similar M&E arrangements to the UFS-CT, but has programme guidelines that require systematic collection and analysis of information on programme implementation and include a household monitoring tool which is used by programme officials conducting monitoring visits. These have been updated into a draft operational manual that sets out monitoring arrangements in more detail. 228. The OPCT also has similar arrangements, but unlike the other programmes has a detailed operational manual that includes a 36 page chapter on Information Management and Monitoring, which is being updated into a joint PWSD-CT/OPCT manual. The 36 page chapter specified very detailed and time-intensive monitoring arrangements, and it appears that these were generally not followed due to inadequate capacity at district level and inadequate demand for M&E information from programme officers. 229. The monitoring procedures of the UFS-CT, PWSD-CT and OPCT are currently being revised as part of the ongoing process to review and improve these procedures. These revised procedures are being documented in a new joint Operations Manual. 230. The HSNP+ M&E arrangements are still in development. HSNP phase 1 M&E arrangements are mainly outsourced to a range of different organisations, coordinated by the HSNP Secretariat (currently based in the MSDNKAL), with very limited roles for government officials, whether at district/sub-county or ministry level. The Secretariat (with DFID) has adequate capacity to manage the coordination of these M&E activities, but all staff are currently not able to interrogate the MIS. The HSNP phase 1 designed explicitly to generate information on the targeting and impacts of the HSNP, so had a well developed targeting and impact evaluation system based around a randomised control trial with three rounds of quantitative survey, qualitative data and a cost effectiveness analysis, conducted by an external firm (Oxford Policy Management). This firm also conducted monitoring of payments and grievances using a separate instrument. A targeting assessment, an impact evaluation baseline and follow up, and quarterly monitoring reports have been submitted to the HSNP Secretariat and DFID. These reports provide high quality measurements of key impact and operational indicators and an independent assessment of targeting effectiveness. The HSNP has an advanced electronic MIS on which the new NSNP programme MISs are being based. The MIS produces information on payments and grievances for each payment cycle, but cannot provide information on the timeliness of payments because individual payments cannot be tied to particular payment cycles, which is problematic for management decisions and will make the monitoring of the DLI on timeliness problematic if not changed. The HSNP struggles to track individual grievances because the grievance system is implemented by a range of sub-contractors. 231. There are some concerns about the adequacy of current M&E individual arrangements for the five programmes and at the Secretariat level to provide regular and accurate monitoring information for the DLIs. These concerns are being addressed through ongoing support from the Bank and other partners. First, the programme MISs need to be upgraded to provide regular and accurate indicators for timeliness of payments (results area 1.4) and complaint resolution (results area 1.5). Second, a PIBS survey of NSNP beneficiaries is required to generate reliable information on whether targeting criteria are followed (results area 1.1), and beneficiaries can name two ways of making a complaint (results area 1.5). Third, capacity at the Secretariat needs to be maintained and built to ensure that accurate information on the other DLIs can be provided, and the PIBS survey adequately managed. 59 6.2.2 PROPOSED ARRANGEMENTS FOR NSNP 232. The proposed NSNP M&E system would be based around elements that are similar to those of the current system, and will be set out in detail in an M&E plan, currently being developed by the government and Bank task team. This is because substantial change would be disruptive and difficult to manage. The Social Protection Secretariat M&E specialist will coordinate the system, supported by the M&E officer, and reporting to the head of the Secretariat. This M&E coordinator would be responsible for compiling reports from the individual programmes to construct NSNP level indicators, conducting narrative reporting for NSNP level indicators, managing NSNP level external surveys (the PIBS Survey), filling out and circulating NSNP reports to government, NIMES and development partners, and ensuring that data are uploaded to Open Data at specified intervals. 233. While it is unrealistic to expect each programme to replicate the four person M&E unit in the CT-OVC, each programme should assign an officer responsible for the coordination of M&E within the programme, and an officer responsible for the MIS. The M&E officers would be responsible for ensuring that the relevant reports are produced and sent to the Secretariat on a timely basis, for coordinating internal spot checks and for managing external surveys and evaluations. These officers would ensure that the required information is emailed to the M&E officer in the Secretariat according to the M&E plan. The MIS officer would ensure that the MIS is kept up to date and data are available to the Secretariat. Sub-county officials would be responsible for providing regular updates to the MIS of the individual programmes, and raising any other issues from beneficiary welfare committees or their own monitoring that require field visits or attention from programme officers. This may require additional staffing at the sub-county level. These updates are either sent via email or hard copy to the programme headquarters in Nairobi, or inputted directly into the MIS (depending on the level of computerisation at the sub-county). Capacity should be built to ensure that this can be automated as much as possible. 234. The key data sources are:  An improved Management Information System (MIS), with information fed in regularly by district (or sub-County) officials via centralised data entry clerks (until the MIS data entry can be decentralised). This would comprise of individual programme MISs and an overall Single Registry as described in section 3.3 above. This would be modelled on the HSNP MIS and will need to involved less time from district officials to be effective.  A Programme Implementation and Beneficiary Satisfaction (PIBS) Survey of NSNP beneficiaries. This survey would be based on the current CT-OVC external monitoring survey and would also be conducted by an external firm annually. The Terms of Reference for the PIBS survey are in development, but would include questions on several results framework indicators and respondents from each beneficiary type. This can be conducted by a Kenyan organisation (as the CT-OVC external monitoring survey currently is), potentially with brief technical support from an international organisation.  More systematised and reduced spot checks by programme officials. Currently programme officials spend a substantial portion of their time conducting spot checks, and these are in their performance agreements, but they lack any systematic process for researching or recording and sharing information. These spot checks produce useful monitoring and learning but could be conducted less frequently and more systematically.  Programme budget data.  External evaluations of the CT-OVC, the HSNP (probably) and other programmes (possibly) that assess impact (covering the key results indicators for each programme and the NSNP), targeting, and provide verification on other operational indicators (such as complaint resolution, knowledge of complaint system, timely payment receipt, and receipt of payment in full). These may need to be conducted by international organisations working together with a Kenyan organisation.  Kenya Integrated Household Budget Survey (KIHBS) data which will allow the programme to measure coverage, targeting and benefit incidence and contribution to households’ 60 consumption.. The M&E officer from the Secretariat and the Bank team are currently in discussion with KNBS about the possibility of including questions on NSNP receipt and eligibility in the KIHBS so that this forms a regular part of the NSNP M&E system. 235. Reporting will take place in accordance with specifications in the M&E plan. Reports will be generated and circulated each payment cycle, quarterly, half yearly and annually. NSNP performance reports will be compiled by the M&E officer in the Social Protection Secretariat, signed off by the Head of the Social Protection Secretariat and shared with programme officials, Department Directors, Permanent Secretaries of the relevant ministries, the Social Protection Secretariat head, the Social Protection Council, and development partners. Capacity to demand and use these reports needs to be built through long-term technical assistance. Selected indicators will also be uploaded to NIMES every six months. Reports will also be made available to Parliament and civil society organisations. A summary of the content of each of these reports is included below:  Bi-monthly reports: These reports will draw on key data from the Single Registry to analyse the performance of each bi-monthly payment cycle with the aim of providing programme managers with information to support implementation. The reports will include information on the number of beneficiaries enrolled into and/or exited from the programmes, the timeliness of payments, complaints lodged and actioned, and budget execution, among others  Quarterly reports: These reports will draw on the Single Registry and also the spot checks carried out by staff of the Secretariats implementing the five cash transfer programmes. These reports will be accompanied by financial reports generated from IFMIS. The aim is to report on key indicators for the Vision 2030 Flagship programmes, which includes the four cash transfer programmes implemented by MGCSD, as required by the Ministry of State for Planning, National Development, and Vision 2030.  Half-yearly reports: These reports will consolidate the information generated in the bi- monthly and quarterly reports and provide further narrative on NSNP inputs, activities and outputs. The aim is to provide information for programme managers, senior government officials and development partners on progress in programme implementation.  Annual reports: These reports will analyse progress against the overall M&E framework. They will draw on data from the programmes’ MISs and Single Registry, spot checks by programme officers, the PIBS survey to monitor programme performance, and evaluations (when these have been reported). These reports will be shared with development partners and will be made public by posting them on the website of the Social Protection Secretariat. They will also inform an annual NSNP learning event. 236. Indicators on outcome will be reported at intervals of two years, or when the evaluations are conducted. These will be reported in evaluation reports submitted by the sub-contracted agencies. The impact indicator (hardcore poverty) will be produced by the KIHBS, expected to take place approximately every five years with the next round of the survey in 2013 and the following round in 2017/2018. However, the KIHBS has not been regular in recent rounds, with the last survey in 2005/2006. 6.3 Capacity Requirements 237. The principal areas of capacity shortfall for the monitoring system are in:  The capacity of district officials to provide regular and reliable information to populate the MIS. This is because they lack the time, materials and management.  The capacity of programme decision-makers to demand and use monitoring and evaluation information. Even where operational or impact reviews are carried out, programme decision- makers often struggle to read and use the reports and results provided.  Capacity to develop, implement/contract out and supervise various evaluations (implementation, process, spot checks, etc.)  The absence of clear reporting formats and procedures that simplify the process of reporting on indicators and reading those reports. 61  The absence of results frameworks and operational manuals in all programmes setting out M&E procedures and key indicators. 238. Capacity should therefore be strengthened in these areas. Capacity at district level is an NSNP wide issue. A functional review should be conducted that includes an assessment of the ability of district officials to complete the required MIS information every payment cycle, and what they would need (additional time, materials or other equipment) to produce this. This may suggest the appointment of additional district staff or the provision of computers and transport allowances. Ways in which the capacity of decision-makers to demand and use information can be strengthened is less obvious, and would probably require long-term technical assistance in the relevant departments, together with improvements in the way in which monitoring information is presented. The development of formats and procedures for monitoring and reporting would be a relatively straightforward and rapid way to improve. 239. In addition, each programme would benefit from a dedicated M&E officer rather than the current sharing of the burden across programme officers. This would make accountability for M&E much clearer and allow some specialisation. The M&E function in the Social Protection Secretariat will need significant support if the current UNICEF-funded consultant leaves. 7 ECONOMIC JUSTIFICATION 7.1 Rationale for Public Provision and Financing 240. Cash transfers have a proven track record internationally in reducing poverty and inequality, increasing household access to food, and improving human development outcomes. These effects are even more significant in countries or areas where poverty is most severe. Analysis of the impact of social pensions has shown that receiving a non-contributory pension reduces the probability of being poor by 18 percent in Brazil and 12.5 percent in South Africa.111 These programmes in Brazil and South Africa have also reduced inequality. Cash grants in South Africa are estimated to have reduced the country’s Gini coefficient by 3 percentage points, while Brazil’s Beneficio da Prestação Continuada de Assistência Social (BPC) and Bolsa Familia combined are responsible for 28 percent of the 2.7 percentage decrease in the Gini coefficient between 1995 and 2004. Such programmes have enabled targeted households to spend more on household necessities such as food, fuel, and housing and to invest more in their children’s health, nutrition, and education112 thus reducing the need to resort to negative coping mechanisms such as selling livestock, taking out high interest loans, or engaging in risky behaviour. 113 These increased investments benefit not only the beneficiaries themselves but also society as a whole, which is a clear justification for the public provision and financing of cash transfers.   7.2 The Programme’s Economic Impact 241. Evidence to date has indicated that significant benefits are likely to accrue to the 3 million beneficiaries expected to participate in the cash transfer programmes of the NSNP. 114 They will smooth the consumption of beneficiary households, increase their human capital, increase their participation in economic life, and improve their psycho-social outcomes. 242. Evaluations of the CT-OVC, HSNP, and OPCT have all demonstrated that households spend a significant proportion of their cash transfer income on food consumption (88 percent in the HSNP 111 Arnold et al (2011) 112 Arnold et al (2011) and DFID (2011) 113 Raghav (2006), Samson et al (2004), and Baird et al (2012) 114 This section draws heavily on the following impact assessments: OPM and IDS (2012), ADR (2011), and OPM (2010). 114 Handa et al (2012) 62 and 71 percent in the OPCT). Other significant expenditure items include health, education, productive investments, debt repayments (in the HSNP), and improvements to housing (in the OPCT). This expenditure on food has increased both the quantity (the number of meals a day) and quality (greater dietary diversity) of the food consumed by households compared to non-beneficiaries (by 15 percent on a relatively simple dietary diversity scale used in the assessment of the CT-OVC). 243. Participating in the cash transfer programmes increased beneficiary households’ use of health and education services more than that of non-beneficiaries. While the OPCT evaluations found a general increase in the use of education services by children in beneficiary households, the evaluation of the CT-OVC found that this impact was largely due to an increase in secondary enrolments and in children attending the appropriate grade for their age, and the HSNP found the biggest effects to have been on grade progression and educational attainment. 244. Cash transfers have enabled beneficiary households to invest in and protect their livelihoods. The OPCT evaluation found beneficiaries investing cash in livestock, farm inputs, and small business ventures while the HSNP evaluation found that beneficiary households had larger livestock holdings than the control group because they had not been forced to sell their livestock to meet their consumption needs (the number of HSNP beneficiary households owning livestock had increased by 5.5 percent, while the number of households in the control group owning livestock had decreased by 6.6 percent). 245. The recent CT-OVC impact evaluation analysed the impact of the programme on health and demographic outcomes of young beneficiaries. The study found that young adults who had benefitted from the programme were less likely to have engaged in unprotected sex and had had fewer sexual partners than those who had not been programme beneficiaries. This in turn means the programme may be playing a role in preventing the spread of HIV. The same study also found that young people, particularly men, who benefitted from the programme were 15 percent less likely to suffer from depressive symptoms and were 19 percent more likely to score above the median on the Hope scale115 than similar non-beneficiaries.116 7.3 World Bank Added Value 246. The World Bank is well positioned to support the Government of Kenya in establishing the National Safety Net Programme. Through its current engagement in the sector, the World Bank has established itself as a knowledge organization that is uniquely positioned to bring international good practice to bear. The World Bank has supported a range of analytical work that has strengthened the overall social protection sector, while increasing the efficiency and effectiveness of individual programmes. Additionally, the World Bank’s convening authority in the sector is well recognized, as is its commitment to supporting government-led, multi-donor processes. This approach was successfully used in the formulation of the National Social Protection Policy, which was recently approved by the Kenyan Cabinet. Finally, World Bank financing to this sector currently enables the provision of safety net support to over 50,000 households through the CT-OVC. World Bank support for capacity building in the sector has also been extensive both at the programme level (most notably within the CT-OVC) and across the sector, particularly in supporting the establishment and strengthening of the Social Protection Secretariat. Existing World Bank financing to the CT-OVC will provide a vehicle to deliver capacity-building support to the NSNP during its first two years of operation.   7.4 Results of Economic Evaluation 247. The 2010 evaluation of the CT-OVC and the 2012 evaluation of the HSNP have demonstrated that cash transfer programmes reduce poverty. This can be seen in the reduction of the number of 115 The Hope Scale is a six item Likert scale measuring children’s perception of their ability to influence their future because they can see ways of achieving their goals and have the motivation to try and achieve them.. 116 Handa et al (2012) 63 people who fall under the national poverty line and who live in extreme poverty. The CT-OVC evaluation indicated a reduction of 16.1 percent in the number of beneficiaries living on less than $1 a day compared to a reduction in the control group of only 2.9 percent. Evidence from the HSNP found that poverty had increased significantly among control households (5.3 percent more households were in the bottom national decile) as a consequence of the severe drought in Northern Kenya at that time. However, there was no equivalent increase among beneficiary households (only 0.1 percent more households were in the bottom decile). 248. The recent CT-OVC impact evaluation looked at the impact of cash transfers on the local economy using a combination of data analysis and simulations. This assessment found that, as the CT-OVC injected cash into local economies (through payments to beneficiaries that in turn were spent in local markets), the demand for goods and services increased. This extended the positive effects of the programme beyond its direct beneficiaries to many non-beneficiaries who were engaged in providing food and other goods and services required by the beneficiary households. The simulations found that the total impact was significantly greater than the amount transferred under the programme in nominal terms, with income multipliers well in excess of 1.0 and in some cases approaching 2.0. These indicate that the transfers have had significant positive spillovers.117   8 INPUTS TO THE PROGRAMME ACTION PLAN 249. The following actions have been identified on the basis of the above analysis. Cross- references are included to allow the reader to see how the analysis informs the proposed action and shows them where to find more detailed recommendations. The list of proposed actions is structured around the three expected results of the Program for Results operation. These are:  To create more robust systems for targeting, registering beneficiaries, making payments, and monitoring performance to strengthen the overall governance of these programmes  To increasingly harmonize the five cash transfer programmes to increase the coherence of the sector  To expand coverage of the five programmes in a coordinated manner that will progressively realize the right to safety net support. 250. The actions listed below will be combined with proposed actions from the Fiduciary Assessment and the Environmental and Social Assessment to form an overall Programme Action Plan. 8.1 To create more robust systems for targeting, registering beneficiaries, making payments, and monitoring performance to strengthen the overall governance of these programmes 251. Strengthened OPCT, UFS-CT, and PWSD-CT programmes (discussed in section 3.2). Significant investment is needed in the systems and procedures for implementing the OPCT, UFS-CT, and PWSD-CT programmes. Work is currently underway to enhance these key aspects of these programmes and to document them in comprehensive programme manuals that are easy to understand. This design process needs to be followed by a comprehensive roll-out, including training, the provision of simple training and guidance materials in Swahili (as well as English) on-the-job guidance and monitoring of implementation to ensure that these key systems and the necessary accompanying resources are in place. 252. Per Capita Programme Generosity (discussed in section 3.4.1). The per capita generosity of each programme has significant implications for the welfare of the programme’s beneficiaries. Per capita programme generosity is affected by inflation levels, which have been high in recent years, and the fact that all five cash transfer programmes currently use a standard transfer value for all households in the programme regardless of their size. There is a need to develop a system to 117 Taylor et al (2012) 64 regularly review transfer levels in relation to inflation in order to provide policymakers with the information that they need to decide whether to adjust the value and by how much. 253. There is also a need to decide whether the transfer amount should vary according to the size of the households. There are several different ways to do this, including: (i) full family targeting; (ii) per eligible member targeting; or (iii) a per-family member or eligible member supplement to a household-level transfer. The government and its development partners should review options for identifying the mechanism that is the most appropriate, easy, and practical to implement in Kenya. 254. Payment Mechanism (discussed in section 3.4.2). Changes in the legislative environment combined with recent developments in international good practice mean that there is a need to review and amend the payment mechanisms that are currently being used by the cash transfer programmes that make up the NSNP. Among the changes that need to be made are the following: (i) new contracts should only be signed with service providers that can comply with the new legislation; (ii) use should be made of mainstream payment infrastructure (such as magstripe cards with PINs), which tends to cost less, to be subject to credible international standards, and to promote financial inclusion; and (iii) beneficiaries should be provided with a bank account that will allow them to take advantage of the benefits of financial inclusion including saving a proportion of their transfer. 255. Timely Transfers (discussed in section 3.4.3). There is substantial evidence that when the delivery of transfers is both timely and predictable, the transfers have a greater positive impact on household welfare than those that are paid late and at unpredictable intervals. There is a need for all programmes to identify and address any bottlenecks in the transfer process in order to increase the predictability of their delivery. Two key bottlenecks have already been identified ‒ the long time needed to undertake reconciliations (because this has to be done manually) and interruptions to the flow of finances from the Treasury as a consequence of the year-end closing of the financial system or cash flow constraints. Programme managers and policymakers should find ways to automate reconciliations and to reclassify cash transfer payments as emoluments. 256. Adequate Budget for Programme Operations (discussed in sections 4 and 5.2). Currently there is no specific budget for the implementation and oversight of the three programmes managed by the Department of Gender and Social Development (DOGSD). Instead, the budget for these activities is mixed in with other financing for this department. There is also no specific budget allocation at the district or country level for the oversight and implementation of programmes managed by the MGCSD. There is a need to modify the chart of accounts to allow such costs to be budgeted (and reported) in sufficient detail to allow proper allocation and tracking. This will involve revising the chart of accounts to include various programme dimensions (such as cash transfers, commissions to payment service providers, and travel costs) and the level of government at which costs are incurred (at the district, county, or national levels). 257. The creation of specific budget lines will then make possible adequate budgeting for oversight and supervision. The UFS-CT, OPCT, and PWSD-CT are still in the early years of their implementation and require significant additional investments in systems and procedures. As a result, it is anticipated that the management costs for these programmes will be high in the initial years of the NSNP before falling in the final years of the first phase. The CT-OVC and the HSNP are more established so their management costs have already fallen. 258. Transparency, Accountability, and Grievance Procedures (see section 3.5). In addition to the improved design of the grievance procedures of UFS-CT, PWSD-CT, and OPCT that will take place as part of the strengthening of these programmes, there is a need to enhance the accountability and grievance systems across the whole of the NSNP. Particular priorities include:  Enhancing awareness of programme and complaints procedures  Improving the documentation of complaints to ensure that they are followed up and using these data to analyse how the programme has responded to complaints and identify areas where there is room for improvement 65  Improving the case management to ensure that complaints are resolved and complainants are informed of the outcome. 259. Monitoring and Evaluation (discussed in section 6). The current monitoring systems and formats outlined in the current guidance material for each programme are complex and are rarely followed. This is partly due to the country’s lack of capacity and the fact that the foundation of the monitoring system is the District Officers who do not currently have the time or the resources to undertake regular monitoring. The planned improvements to the MISs and the development of the single registry will facilitate the analysis and reporting of monitoring information, but human and financial resources are still needed at the district level to gather basic data. There is a need both to simplify the current monitoring and evaluation system and to enhance the capacity to administer it. 260. In addition, there is a critical need to increase the demand for monitoring information and its use by programme decision-makers. The development of simple reporting formats will facilitate this, but further work is needed to help the MGCSD and other government stakeholders to use evidence and lessons learned from monitoring in programme decision-making, thus increasing transparency and accountability. Core indicators will also be reported by the Social Protection Secretariat to the government’s mainstream monitoring system (NIMES) and on the open data website. Key activities and results will also form part of the performance contract between the MGCSD and the central government. 8.2 To gradually harmonize the five cash transfer programmes to increase the coherence of the sector 261. Cross-Learning between Programmes (see sections 3.2.1 and 4.2.1.1). A major advantage of the NSNP approach is that it increases the potential for programmes to learn lessons from each other’s experience. This should be put into operation immediately so that the lessons learned can inform any refinements to the cash transfer components prior to the launch of the NSNP next year. Some ways in which this can be put into practice include:  Strengthening the OPCT, the PWSD-CT, and the UFS-CT as mentioned above  Improving geographical targeting in all programmes implemented by the MGCSD (described in more detail below)  Building on the positive experiences of the CT-OVC and the HSNP in establishing and working with local committees made up of beneficiaries. These committees can play a key role in the forwarding of complaints as well as communication on a range of programme issues.   262. Functional Assessment (see section 4.4). The future institutional arrangements of the NSNP should facilitate this cross-learning, enhance efficiency, and improve the quality of programme implementation. A functional review of the NSNP will be conducted in mid-2013 in order to inform the design of the NSNP’s institutional arrangements. This review needs to consider:  The overall programme organogram and how the programmes managed by the MGCSD will be coordinated among themselves and with the HSNP and the Contingent Financing Component  How to ensure adequate staffing of the programme including within the Social Protection Secretariat, within the various departments at the national level that are responsible for programme implementation, and at the county and district levels  Any need for short-term or medium-term technical assistance to support the implementation units  The consequences of devolved governance for programme implementation and the potential for increased competing demands on district and county level staff. 263. Enhance Systems to Allow Programme Consolidation (discussed in sections 4.4). In addition to the functional review, the government will prepare a plan for how to further harmonize cash transfer programmes in order support cross-learning, enhance efficiency, and improve the quality of 66 programme implementation. It will be important to allow the government to drive this process to ensure that it has full ownership of the resulting plan. 8.3 To expand coverage of the five programmes in a coordinated manner that will progressively realize the right to safety net support 264. Geographical Targeting and Resource Allocations (see section 3.2.1.1). There is a need for better evidence on the geographical distribution of vulnerability and poverty to inform decisions about which locations to prioritize and how best to allocate resources at the district and location levels. A clear justification for how resources are allocated should mitigate political pressure to spread resources evenly throughout the country. Discussions between the government and development partners on how to analyse available data are already underway. The plan should take into consideration the latest data on poverty and vulnerability118, financing available and the levels of scale up planned for each programme119, current levels of coverage by programme by location, how best to use limited implementation capacity, and agreements on how differing programmes will work together to expand coverage.   9 TECHNICAL RISK RATING 265. The technical risk rating for this project is medium. The major technical risks arise from: (i) the policy context; (ii) the need for the government to continue to demonstrate strong leadership to drive forward the proposed programme; (iii) ongoing risks of error, fraud, and corruption; (iv) the likely forthcoming ministerial restructuring along with the transition process for devolution; and (v) the generally weak capacity in the implementing line ministries for which a number of mitigating activities are already ongoing. Table 18 below summarizes these risks and some of the main mitigation measures. Table 18: Technical Risk Assessment and Mitigating Measures Description of Risk Mitigation Measures Delay in passing National Social Protection Policy: The National Social Protection Policy was recently While the National Social Protection Policy creates an approved by the Cabinet. While it currently enjoys important framework for the sector, the broad policy broad-based political support, the policy may not be directions have already been well established in the passed by Parliament before the coming election. This 2010 Constitution, which established the right to social would weaken the framework for the sector in the short protection, and in Vision 2030, which committed the term and increase the risk that the election may lead to government to increasing financing for safety nets. new political priorities that could undermine the Additionally, there is strong political interest in progress made to date on the policy. expanding the coverage of safety nets in both the Parliament and the Executive, and the government continues to demonstrate strong leadership in this area. Weak Leadership due to lack of attention during election The two implementing Ministries (the MGCSD and the While the proposed National Safety Net Programme MSDNKAL) have demonstrated strong leadership on will put into operation the policy vision for this sector, cash transfers in Kenya. The Office of the Prime strong government leadership will be required to Minister has taken on a sector-wide coordinating role, realize this vision. This is because of the institutional with the aim of facilitating the establishment of the reforms that this process will require and the number National Safety Net Programme with strong support of stakeholders involved. Weak government leadership from Treasury. At the same time, the Prime Minister or the preoccupation of senior decision-makers with has indicated his strong commitment to this agenda. the upcoming election could undermine the proposed programme. 118 At present this is the 2005/06 Kenya Income and Household Budget Survey and the 2009 Census 119 An NSNP Medium Term Expenditure and Financing Framework (MTEFF) has been developed and is currently under review. The final version of the MTEFF will be annexed to the governments programme document for the NSNP. 67 Need to minimize error, fraud, and corruption: It is inevitable to have some degree of error, fraud, and A number of efforts are already being made to improve corruption (EFC) in any cash transfer programme. the prevention, detection, deterrence, and monitoring Examples of EFC in the NSNP include beneficiary of error, fraud, and corruption. These include (i) efforts households who do not meet the eligibility criteria (for to strengthen the targeting systems used by the three example, do not meet the age requirement of the OPCT weakest programmes; (ii) the upgrading of the or are not taking care of an OVC for the CT-OVC), complaints and grievance procedures throughout the allegations of discrimination, corruption, and nepotism, NSNP; (iii) investments in programme MISs and a suggestions of ghost beneficiaries in UFS-CT, some single registry that will improve payroll controls, thus beneficiaries participating in more than one cash making it simpler to prevent double-dipping and to transfer programme, and changes in beneficiary minimize the potential for ghost beneficiaries; and (iv) households not reflected on the payroll. the introduction of two-factor authentication for all payments in the NSNP. Unknown consequences of government restructuring: It is likely that any ministerial-level restructuring will It is likely that there will be a restructuring of involve the wholesale movement of departments. A ministries following the March 2013 election. This functional review has been planned that will assess the may result in a significant reduction of ministries, with programme’s capacity needs against current existing likely implications for both the MSDNKAL and the capacity. This functional review will need to consider MGCSD. the question of competing demands and make There are also likely to be implications for the recommendations appropriately. implementation and oversight of the cash transfer programmes as a consequence of the moves towards devolution. This will include increased scrutiny by county governors but also potentially competing demands on county and district staff who will have to report on both county and national-level functions. Furthermore, there is likely to be a period of confusion during the transition process that will affect all aspects of government service delivery. Capacity constraints in the sector: Overall capacity in the sector remains low, and this is The government has commissioned an options paper to compounded by significant fragmentation. explore how to strengthen the implementation capacity Implementation capacity: The five cash transfer of the four cash transfer programmes housed in the programmes are implemented by separate secretariats MGCSD. Development partners, including the World housed in two different ministries with distinct Bank, are supporting capacity-building efforts in many implementation arrangements at county-level. This of these Secretariats, including the Social Protection means that the number of staff and the skill mix within Secretariat. In this case, the programme itself will each secretariat is less than is required for effective significantly strengthen the Social Protection programme implementation. Secretariat by giving it a clear mandate to oversee the Coordination and oversight capacity: The sector is proposed programme. Existing World Bank financing coordinated by the Social Protection Secretariat, but to the CT-OVC will deliver capacity-building support this body is also relatively weak as it has only recently to the NSNP during its first two and a half years of been established by the NSNP. operation.   10 INPUTS TO THE PROGRAMME IMPLEMENTATION SUPPORT PLAN 266. This operation marks a significant departure from the World Bank’s traditional ways of supporting the Government of Kenya. The focus on results rather than inputs and activities is a new approach for the government, and it will need support to meet the demands of this new approach. In addition, this programme brings together a number of development partners, government departments and ministries, and new actors (such as the NDMA) in ways that will require a more harmonized approach by donors and greater coordination and harmonization between programmes and among stakeholders. 267. Regular supervision and operational support reviews and meetings are already taking place under the CT-OVC. These include the six-monthly Joint Review and Implementation Support Missions, a fortnightly Technical Working Group meetings (attended by development partners and the CT-OVC secretariat), and Quarterly Programme Management Oversight Committees (attended by 68 Technical Working Group members and other relevant MGCSD staff and chaired by the Secretary of Children’s Services). A similar pattern of joint supervision and review along with regular technical and programme management meetings will need to be established for the NSNP as a whole in order to support the government through the transition process and address any emerging risks. Table 19: Proposed Schedule of Meetings and Operational Support Reviews Joint Review and Supervision One Mission held programme-wide to include all key stakeholders Missions including the MGCSD and its implementation departments, the HSNP Secretariat and its responsible ministry, and the NDMA Quarterly (or Four-monthly) One meeting to include all key stakeholders including the Programme Management MGCSD and its implementation departments, the HSNP Oversight Meetings Secretariat and its responsible ministry, and the NDMA Technical Level Meetings  Specific regular meetings concerning programmes managed by the MGCSD  Specific regular meetings concerning the HSNP  Regular meetings during the start-up of the contingent financing component, followed by ad hoc meetings when a response is triggered Specific Issue Working Groups Cross-programme specific issue working groups will be established as necessary. Possible areas include a Single Registry Working Group (already functioning) and a Payments Working Group (agreed but not yet operational). Further working groups may need to be identified through the coordination and harmonization agenda.   69 REFERENCES ADR (2011), “Impact of the Older Persons Cash Transfer (OPCT) Program in Kenya,” Agency for Development Research, Kenya Alviar, C. (2009), “Evidence Based Policy Analysis to Deliver Results for Children: Final Policy Paper” UNICEF Learning Programme, UNICEF, Kenya Amuyunzu-Nyamongo, M. (2012), “Draft Environmental and Social Assessment for the Program for Results on Social Safety Nets in Kenya”, African Institute for Health and Development, Kenya Arnold, C., T. Conway, and M. Greenslade (2011) “DFID Cash Transfers Literature Review”, Policy Division, Department for International Development, London, UK Baird S., C. McIntosh, and B. Ozler (2011), “Cash Or Condition? Evidence from a Cash Transfer Experiment,” The Quarterly Journal of Economics, 1–44, Oxford University Press Baird S.J., R.S. Garfein, C.T. McIntosh, and B. Ozler (2012), “Effect of a Cash Transfer Programme for Schooling on the Prevalence of HIV and Herpes Simplex Type 2 in Malawi: a Cluster Randomised trial,” Lancet. 2012 April 7; 379(9823):1320-9. Epub 2012 February 15 Barr, D. (2012), “Analysis of Error Fraud and Corruption: Kenya Cash Transfer Programmes,” Unpublished Calder, R., R. Chirchir, C. Cosgrove, S. Kidd, J. Sandford, J. Sibley and E. Wylde (2011), “Review of Targeting in the CT-OVC Programme,” Development Pathways, London, UK Caldés, N., D. Coady and J. Maluccio (2004) “The cost of poverty alleviation transfer programmes: a comparative analysis of three programmes in Latin America”, FCND discussion paper no. 174, February, International Food Policy Research Institute, Washington, D.C. Castañeda, T. and K. Linder with B. de la Brière, L. Fernandez, C. Hubert, O. Larrañaga, M. Orozco, and R. Viquez (2005), “Designing and Implementing Household Targeting Systems: Lessons from Latin America and the United States,” Social Protection Discussion Paper Series No. 0526, World Bank, Washington D.C. Chirchir, R. and S. Kidd (2011) “Scoping Study for a “Single Registry” MIS for Kenya’s Social Protection Programmes,” Development Pathways, London Concern (2012), “Evaluation of Concern Kenya’s Mombasa County Urban Food Subsidy Cash Transfer Implementation,” Concern Kenya, Nairobi Degol Hailu , M. Medeiros, and P. Nonaka (2008), “Legal Protection for Cash Transfers: Why We Need It,” Poverty in Focus Number 15, August 2008, International Poverty Centre for Inclusive Growth, Brasilia, Brazil DFID (2011), “Cash Transfers Evidence Paper,” Policy Division, Department for International Development, London, UK Expert Working Group (2012), “Management Information Systems Guidelines for Social Protection Programmes in Kenya,” Unpublished Fiszbein, A. and N. Schady (2009), “Conditional Cash Transfers for Attacking Present and Future Poverty,” World Bank Policy Research Report, World Bank, Washington D.C. Gilligan, D., J. Hoddinott, N.R. Kumar, and A.S. Taffesse (2009). “An Impact Evaluation of Ethiopia’s Productive Safety Nets Programme,” International Food Policy Research Institute (IFPRI), Washington D.C. Government of Kenya (2012), “Performance Contract between the Government of Kenya and the Ministry of Gender, Children, and Social Development for the Period 1st July, 2012 – 30th June, 2013” Handa, S., C. Huang, K. Kilburn, C. Halpern, A. Pettifor, M. Rosenberg, and H. Thirumurthy (2012), “Impact of the Kenya CT-OVC on the Transition to Adulthood,” University of North Carolina, Chapel Hill, NC 70 Hobson, M. and L. Campbell (2012), “How Ethiopia’s Productive Safety Net Programme (PSNP) is responding to the current humanitarian crisis in the Horn,” Humanitarian Exchange Magazine, Issue 53 (http://www.odihpn.org/ humanitarian-exchange-magazine/issue- 53/how-ethiopias-productive-safety-net-programme-psnp-is-responding-to-the-current- humanitarian-crisis-in-the-horn) Kenya CT-OVC Evaluation Team (2012a), “The Impact of the Kenya CT-OVC Program on Human Capital,” Journal of Development Effectiveness, Vol.4(1): 38-49, 2012 Kenya CT-OVC Evaluation Team (2012b), “Impact of the Kenya CT-OVC on Household Spending,” Journal of Development Effectiveness, Vol.4(1):9-37, 2012. Kimetrica, (2010a), “External Monitoring Guidance December 2010,” Kimetrica International Ltd, Nairobi Kimetrica (2010b), “Spot Check Report: 14 July 2010 (revised 24 November 2010), External Process Monitoring System for the CT-OVC Programme in Kenya,” Kimetrica International Ltd, Nairobi Kimetrica (2012a), “Field Report, Support to the Design of the National Safety Net Program: Strengthening the Older Persons Cash Transfer programme (OPCT) and the Programme for Persons with Severe Disability (PWD),” Kimetrica International Ltd, Nairobi Kimetrica (2012b), “Review of the OPCT and PWSD-CT Safety Net Programs: Older Persons Cash Transfer Persons with Severe Disabilities Cash Transfer,” Presentation to Technical Design Team Meeting 28th-30th November 2012, Kimetrica International Ltd, Nairobi Kimetrica (2012c), “Spot Check Report for March and April 2012, External Process Monitoring System for the CT-OVC Programme in Kenya,” Kimetrica International Ltd, Nairobi Kimetrica, (2012d), “Spot check report October 2011 to January 2012, External Process Monitoring System for the CT-OVC Programme in Kenya,” Kimetrica International Ltd, Nairobi Kimetrica (2013), “Operational Review of the OPCT and PWSD CT Safety Net Programs,” Kimetrica International Ltd, Nairobi KNBS (2007), “Basic Report on Well-being in Kenya based on Kenya Integrated Household Budget Survey - 2005/06,” Kenya National Bureau of Statistics, Nairobi KNBS (2010), “Kenya Census 2009,” Kenya National Bureau of Statistics, Nairobi KNBS and ICF Macro (2010), “Kenya Demographic and Health Survey 2008-2009,” Kenya National Bureau of Statistics, Nairobi, and ICF Macro, Calverton, Maryland MGCSD (2010), “Capacity Building Strategy for Social Protection for the Ministry of Gender, Children and Social Development,” Ministry of Gender, Children, and Social Development, Nairobi MGCSD (2011), “Kenya National Social Protection Policy,” Ministry of Gender, Children, and Social Development, Nairobi MGCSD (2012), “Payroll and MIS Technical Review Report,” Ministry of Gender, Children, and Social Development, Nairobi MGCSD (2013), “Presentation by the Ministry of Gender, Children and Social Development on the Second Medium Term Plan 2013-2017”, Presentation to Pre-Appraisal Mission Retreat 12th-15th February 2013, Ministry of Gender, Children, and Social Development, Nairob MacAuslan, I. (2012), “Oxfam GB Emergency Food Security and Livelihoods Urban Programme Evaluation: Final Report,” Oxford Policy Management (OPM), Oxford Manley J., S Gitter, V. Slavchevska 2011, “How Effective are Cash Transfer Programmes at Improving Nutritional Status?”, Towson University, Department of Economics, Working Paper Series. Ministry of Finance (2012), “Social Protection, Culture and Recreation Sector MTEF 2012/13 – 2014/15 Sector Report,” Nairobi Ministry of State for Planning, National Development, and Vision 2030 (2012), “Kenya Social Protection Sector Review,” Nairobi 71 MoARD (2010), “Programme Implementation Manual (Revised),” Ministry of Agriculture land Rural Development, Addis Ababa, Ethiopia National Council for Law Reporting (2010), “The Constitution of Kenya,” Nairobi NCAPD and KNBS (2008), “Kenya National Survey for Persons with Disabilities: Main Report,” National Coordinating Agency for Population and Development, Nairobi, and Kenya National Bureau of Statistics, Nairobi Ndirangu, L. (2010), “Poverty, Risk and Vulnerability in Kenya,” Kenya Institute for Public Policy Research and Analysis (KIPPRA), Nairobi Ndung’u, M (2012), “Overview of the Government Budget Cycle, and Influencing Allocation to Social Protection Sector and Cash Transfer Programmes through the budget process” Briefing paper prepared for the World Bank, Nairobi Office of the Prime Minister (2012), Report on Evaluation of the Performance of Public Agencies for the Financial Year 2010/2011, Performance Contracting Department OPM (2010), “Cash Transfer Programme for Orphans and Vulnerable Children (CT-OVC), Kenya: Operational and Impact Evaluation, 2007-2009,” Oxford Policy Management (OPM), Oxford OPM (2012a), “Consolidated Operational Monitoring Report, Kenya Hunger Safety Net Programme: Monitoring and Evaluation Component,” Oxford Policy Management (OPM), Oxford OPM (2012b), “Mombasa Urban Food Subsidy Programme: Draft Programme Document,” Oxford Policy Management (OPM), Oxford OPM and IDS (2011), “HSNP Targeting Effectiveness Evaluation Report, Kenya Hunger Safety Net Programme: Monitoring and Evaluation Component,” Oxford Policy Management (OPM), Oxford and Institute of Development Studies (IDS), Brighton OPM and IDS (2012a), “Qualitative Impact Evaluation Report: 2009/10 to 2010/11, Kenya Hunger Safety Net Programme: Monitoring and Evaluation Component,” Oxford Policy Management (OPM), Oxford and Institute of Development Studies (IDS), Brighton OPM and IDS (2012b), “Quantitative Impact Evaluation Report: 2009/10 to 2010/11, Kenya Hunger Safety Net Programme: Monitoring and Evaluation Component,” Oxford Policy Management (OPM), Oxford and Institute of Development Studies (IDS), Brighton Pinney, (2012), “Provisional monitoring and evaluation framework for HSNP+”, Unpublished Promin Consultants and Development Impact Consulting, (2012), “External monitoring for the CT- OVC program, Kenya. Draft report, phase 1,” Promin Consultants and Development Impact Consulting, Nairobi Pulver, C. (2012), Strategic Assessment of Payment Services for the Kenya National Safety Net Programme-For-Results, Unpublished Raghav, Gaiha (2006), “The Maharastra Employment Guarantee Scheme,” Policy Brief 6, Inter- Regional Inequality Facility, Overseas Development Institute, London, UK Republic of Kenya (2013) “Drought Risk Management and Ending Drought Emergencies: Medium- Term Plan, 2013-17”, Final Draft, February 2013. Nairobi: National Drought Management Authority Samson, M., I. van Niekerk, and K. MacQuene (2010), “Designing and Implementing Social Transfer Programmes: Second Edition,” Economic Policy Research Institute (EPRI), Cape Town, South Africa Samson, M., U. Lee, A. Ndlebe, K. MacQuene, I. van Niekerk, V. Gandhi, and T. Harigaya (2004), “The Social and Economic Impact of South Africa’s Social Security System.” Commissioned by the Economics and Finance Directorate, Ministry of Social Development, Republic of South Africa Sandford, J. (2012), “Addressing Vulnerabilities Post-2014: What are the Options?” A Joint Government-Development Partner Position Paper, Stakeholder Analysis on Responses to Vulnerabilities in Ethiopia Post-2014, Unpublished 72 Solomon Asfaw, Benjamin Davis, Josh Dewbre, Giovanni Federighi, Sudhanshu Handa and Paul Winters, (2013)“The Impact of the Kenya CT-OVC Programme on Productive Activities and Labour Allocation.” Paper prepared for the “From Protection to Production Project” (http://www.fao.org/economic/ptop/en/), supported by DFID. Taylor, J.E., J. Kagin, and M. Filipski (2012), “Evaluating General Equilibrium Impacts of Kenya’s Cash Transfer Program for Orphans and Vulnerable Children (CT-OVC),” Department of Agricultural and Resource Economics, University of California The Standard, (September 4, 2009), “High Court outlaws 210 districts created by Moi and Kibaki since 1992,” Nairobi UN-HABITAT (2010), “Global Urban Indicators Database,” United Nations Human Settlements Programme, Nairobi UNICEF (2010), “Considerations on the benefit transfer level of the CT-OVC,” Presentation made to CT-OVC Joint Review Mission, November 15, 2010, UNICEF, Nairobi USAID (no date). “Food Security Framework, Key Population Groups,” US Agency for International Development http://www.fews.net/ml/en/info/Pages/fmwkfactors.aspx?gb=ke&l=en Van Stolk, C. and E. Tesliuc (2010), “Toolkit on Tackling Error, Fraud, and Corruption in Social Protection Programs,” Social Protection Discussion Paper No 1002, World Bank, Washington D.C. White, P. (2011) “Economic appraisal of LEAP Support and Expansion Programme, Ghana”, prepared for Business Case drafted by A. McCord, T. Awori and P. White, Overseas Development Institute, London and DFID Ghana, May White, P. and F. Ellis (2012), “Ethiopia’s Productive Safety Net Programme, 2010-2014 – A Value for Money Assessment,” International Development, University of East Anglia, Norwich, UK White, P. and A. McCord (2006) “Cost comparison of cash, food and agricultural input transfer schemes in Malawi and Zambia”, unpublished paper for DFID, 30 January World Bank (2006), “Kenya Inside Informality: Poverty, Jobs, Housing, and Services in Nairobi’s Slums.” World Bank, Washington D.C. World Bank (2008), “Kenya Poverty and Inequality Assessment, Volume I: Synthesis Report,” Poverty Reduction and Economic Management Unit, Africa Region, World Bank, Washington D.C. World Bank (2012), “Kenya National Safety Net Program for Results: Program for Results Concept Note,” Africa Region, World Bank Washington D.C. 73 Annex 1: RESULTS FRAMEWORK FOR NATIONAL SAFETY NET PROGRAMME120 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility Impacts: Reduced poverty and vulnerability in Kenya 2013; 1 Hardcore poverty headcount % 19.1 (2005/06) NA NA NA NA KIHBS KNBS 2017/18 Outcome 1: Improved beneficiary well-being Net change in beneficiary Evaluations of 274 (CT- 2013, 2014, 2015 household monthly per adult CT-OVC and Programmes 2 Ksh OVC); 150 150 150 150 (CT-OVC); equivalent consumption HSNP, procure 0 (HSNP) HSNP unknown expenditure possibly others Evaluations of Net change in beneficiary 0.82 (CT- 2013, 2014, 2015 CT-OVC and Programmes 3 household dietary diversity Score OVC); 0.5 0.5 0.5 0.5 (CT-OVC); HSNP, procure score 0.52 (HSNP) HSNP unknown possibly others 4.4 (CT-OVC Evaluations of Net change in % beneficiary 6-17 year olds 2013, 2014, 2015 CT-OVC and Programmes 4 household members aged 4-17 % currently 2 2 2 2 (CT-OVC); HSNP, procure currently attending school enrolled); HSNP unknown possibly others 0 (HSNP) Evaluations of 2013, 2014, 2015 Net change in morbidity rates 0 (CT-OVC); CT-OVC and Programmes 5 % 0 0 0 0 (CT-OVC); in beneficiary households 0 (HSNP) HSNP, procure HSNP unknown possibly others Net change in mean self- Evaluations of 2013, 2014, 2015 esteem score of decision- CT-OVC and Programmes 6 Score NA NA NA NA NA (CT-OVC); makers in beneficiary HSNP, procure HSNP unknown households possibly others 120 Disbursement Linked Indicators are underlined 74 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility Outcome 2: Improved beneficiary resilience 0 (CT-OVC, most asset Evaluations of Net change in beneficiary 2013, 2014, 2015 indicators); CT-OVC and Programmes 7 household asset ownership % 0 0 0 0 (CT-OVC); 11 (HSNP, HH HSNP, procure index HSNP unknown owning any possibly others livestock) Evaluations of Net change in % beneficiary 2013, 2014, 2015 0 (CT-OVC); CT-OVC and Programmes 8 households reporting operating % 0 0 0 0 (CT-OVC); 0 (HSNP) HSNP, procure an income generating activity HSNP unknown possibly others Output 1: The most vulnerable and poorest households are enrolled TBC (CT- % NSNP beneficiaries who PIBS Survey; OVC); SP Secretariat to 9 conform to programme % 70% 78% 85% 93% Annual confirm in 84.5 (UFS-CT); procure targeting criteria evaluations 89 (HSNP) M&E officer in Number of households on the House- 10 298,900 298,900 390,740 441,303 519,813 Bi-monthly MIS social protection NSNP payroll* holds secretariat Expansion plan for NSNP, informed by poverty and Expansion vulnerability criteria, agreed, M&E officer in plan; MTEF; 11 ranking localities by poverty Yes/no No Yes Yes Yes Yes Annual social protection Printed and specifying order of secretariat Estimates intervention and number of beneficiaries Number of new households M&E officer in enrolled in the NSNP House- 12 0 0 65,000 130,000 235,000 Bi-monthly MIS social protection according to agreed expansion holds secretariat plan and paid for by GOK 75 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility Output area 2: Beneficiaries receive appropriate, reliable and accessible payments % beneficiaries for whom payments are made M&E officer in MIS; PSP 13 electronically through payment % 41% 50% 60% 75% 90% Bi-monthly social protection contracts services providers using two secretariat factor authentication MIS payroll % all payments disbursed to information; M&E officer in 14 payment service providers on % 26% 40% 50% 60% 70% Bi-monthly planned social protection time payment secretariat window % beneficiaries travelling 67% (CT- SP Secretariat to 15 more than 6km (HSNP 20km) % OVC); 50% 40% 30% 20% Annual PIBS survey procure to the nearest paypoint 6.7% (HSNP) % beneficiaries reporting SP Secretariat to 16 paying bribe to receive % 7% (HSNP) <5% <5% <5% <5% Annual PIBS survey procure payment in last payment cycle M&E officer in 2000; 2000; 2000; 2000; 2013 2000; CPI (from Social Protection 17 Real value of transfer HSNP HSNP HSNP HSNP Annual Ksh HSNP 2300 KNBS) Secretariat to 2300 2300 2300 2300 calculate Output 3: Citizens are able to appeal and complain to improve results % people who know SP Secretariat to 18 programme objectives and % 37% (HSNP) Annual PIBS survey procure entitlements 38.6% (CT- % program beneficiaries who OVC); SP Secretariat to 19 can name two means of % 30% 50% 47.50% 65% Annual PIBS survey 15% (OPCT/ procure making a complaint PwSD-CT) 76 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility OMs; MISs; letter on staff Local Local and hotlines; The complaints and grievance M&E officer in National National and and summary 20 mechanism is functional for all Yes/no No Annual social protection level level national national complaints five cash transfer programmes secretariat level level report; service charters; PIBS Survey 87% (CT-OVC M&E officer in 21 % complaints actioned % external 50% 60% 70% 80% Bi-monthly MIS social protection mechanism) secretariat % complaints actioned within M&E officer in 22 timeframe set out in % NA 50% 60% 70% 80% Bi-monthly MIS social protection operational manual secretariat Output 4: Monitoring and learning system functioning Bi-monthly, quarterly, half- Six bi-monthly reports per year;Four M&E officer in yearly and annual reports Reports;SP 23 Number NA quarterly reports per year;Two half yearly Annual social protection available in accordance with Sec website reports per year;One annual report per year. secretariat defined monitoring framework Specified indicators are NIMES and M&E officer in available in NIMES and on 24 Yes/no NA All data available and updated each year Annual Open data social protection Open Data and are updated websites secretariat annually MIS/SR Single registry is operational systems M&E officer in with program MISs using 25 Yes/no No Yes Yes Yes Yes Annual assessment; social protection agreed standards for internal internal audit secretariat payroll controls reports; OMs Output 5: Cash transfer sector is harmonised and government managed All financing for the NSNP is M&E officer in on the GoK budget and GoK Printed 26 Yes/no Yes Yes Yes Yes Yes Annual social protection provides financing to the estimates secretariat HSNP 77 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility M&E officer in % NSNP spending Printed 27 % 39% 41% 56% 66% 79% Annual social protection government financed estimates secretariat Letter from Strategy for consolidating cash MGCSD M&E officer in 28 transfer programs in MGCSD Yes/no No Yes Yes Yes Yes Annual adopting social protection adopted strategy; secretariat strategy Output 6: Programs are responsive to shocks NDMA and A system for scaling up the HSNP OMs; NSNP as part of the national NDDCF M&E officer in disaster risk management regulations social protection 29 Yes/no No No Yes Yes Yes Annual system has been created with and MOU secretariat to agreed levels of GoK with NDMA; procure contingency financing PSP arrangements Amount committed in GoK National Drought and Disaster M&E officer in USD50 USD50 USD50 Printed 30 Contingency Fund budget line Ksh USD0m USD0m Annual social protection m m m estimates on expanding NSNP programs secretariat in response to agreed triggers Number of households per No targets, but M&E officer in year paid through the NSNP House- number to be 31 NA NA Half yearly MIS social protection out of NDDCF (either existing holds compared to secretariat or additional beneficiaries) population in need Average time taken between MIS; early M&E officer in early warning triggers and Target to be 32 Days NA NA Annual warning social protection citizens receiving NDDCF developed in OMs trigger secretariat cash through NSNP 78 Unit of Target Means of verification Indicators Baseline measure 2014 2015 2016 2017 Frequency Sources Responsibility Input area 1: resourcing is adequate, affordable and sustainable 33 Total NSNP expenditure* Ksh 10.76bn 12.17bn 14.30bn 17.48bn Bi-monthly IFMIS Programme finance officers Amount spent on NSNP by IFMIS; GDP to send to M&E 34 % 0.1% 0.001 0.14% 0.18% 0.26% Annual GoK as % of GDP from KNBS officer in SP Secretariat to compile NSNP operational costs as % 35 % 15% 15% 15% 15% Annual IFMIS of NSNP total costs Input area 2: Staff and organisational capacity adequate for efficient and effective implementation MIS; Number of beneficiaries per House- MGSCD, 36 Functional review to define targets Annual NSNP officer holds NDMA payroll PIBS Survey % NSNP-related offices with or 37 % Functional review to define targets Annual adequate capacity for NSNP administrative M&E officer in data social protection MGCSD and secretariat to % NSNP officers that have NDMA calculate 38 received NSNP-related % Target to be developed in OMs Annual payrolls; training over the last two years Administrative data % of NSNP positions filled MGCSD and 39 according to the functional % Functional review to define targets Annual NDMA review payrolls 79