Emergency Food Security and Livelihoods Project in Amhara and Oromia regions

By Sarah Coll-Black and Matt Hobson

Sarah Coll-Black is a Social Protection Specialist working with the World Bank in Ethiopia and Kenya. She has been involved with Ethiopia's Productive Safety Net Programme (PSNP) since 2007, initially as a member of the PSNP Donor Coordination Team and then with the World Bank PSNP Task Team. She holds an MPhil from the Institute of Development Studies, University of Sussex.

Matt Hobson is the Donor Coordinator for the PSNP. Previously, he worked for Save the Children as Head of Hunger Reduction in Ethiopia, as well as in Somalia, Sri Lanka, Afghanistan and London. Prior to this, Matt worked as a human rights lawyer in private practice.

This article is based on a longer publication entitled 'Designing and Implementing a Rural Safety Net in a Low Income Setting: Lessons Learned from Ethiopia's Productive Safety Net Program 2005-2009' by Will Wiseman, Julie Van Domelen and Sarah Coll-Black.

For generations, many rural Ethiopians have experienced significant periods when they were unable to meet their basic food needs. For over 20 years, the main response to this situation was food aid to ensure basic survival. Ethiopia has received more emergency support per capita than any other Sub-Saharan African nation, with appeals rising to 14.5 million people in some years. On average, 700,000 metric tons (MT) of food aid per year have been delivered in Ethiopia since 1990.

Serious deficiencies with the emergency food aid system were increasingly clear by the late 1990s. Despite high levels of food aid, with each emergency rural households further depleted their assets and found themselves increasingly vulnerable to even the most marginal livelihood shock. Analyses of these issues revealed that the timeliness of food aid deliveries was a serious deficiency in the system. There was also growing concern of the potentially negative effects of such large volumes of food aid on local food markets.

By late 2003, a technical group of the New Coalition for Food Security¹ had developed a proposal to reform the emergency appeal system in favour of an integrated approach to reducing vulnerability and food insecurity. This approach brought together under a single umbrella a range of initiatives implemented by the Government, donors and non-governmental organisations (NGOs).

For the Government, the objectives of the New Coalition for Food Security (which was led by the Prime Minister, Ato Meles Zenawi) would be largely met through implementation of the Food Security Programme (FSP). This consisted of three strategic pillars:

1. Resettling households from unsustainable and environmentally degraded lands
2. Developing a safety net for chronically food insecure households
3. Supplying agricultural and financial services to food insecure households to promote their graduation out of food insecurity.

Of these three pillars, donors embraced the safety net, the subsequent design of whichwas supported by a safety net donor group that engaged intensely with Government. The Productive Safety Net Programme (PSNP) was designed from late 2003 to the end of 2004 and was launched at scale in February 2005. It was not meant immediately to replace the emergency appeal system; instead the emergency appeal system would be gradually reduced over the course of two years.

PSNP Operations (2005-2009)

The objective of the PSNP (2005-2009) was to ensure food consumption and prevention of asset depletion for rural food insecure households in a way that stimulates markets, improves access to services and natural resources and rehabilitates and enhances the natural environment.

The PSNP provides cash and/or food transfers to households through two mechanisms:

1. Chronically food insecure households with able-bodied adults receive a transfer for their participation in public work
2. Chronically food insecure households who cannot provide labour to public works and have no other means of support are provided an unconditional transfer. Direct support beneficiaries include, but are not limited to, orphans, pregnant and nursing mothers, people living with disabilities, the elderly chronically ill individuals and femaleheaded households that are labour poor.

PSNP stakeholders

The Ministry of Agriculture and Rural Development (MOARD) is responsible for the management of the PSNP, with the Disaster Risk Management and Food Security Sector (DRMFSS) responsible for overall programme coordination.

The Food Security Co-ordination Directorate is one of two Directorates that report to the DRMFSS and is the main implementing Directorate for the PSNP. The Early Warning and Response Directorate (EWRD), which is the second Directorate under DRMFSS, provides accurate and timely early warning information for the PSNP Risk Financing Mechanism (RF) and ensures linkages between PSNP RF and other humanitarian response activities. The Natural Resource Management Directorate (NRMD) within MOARD is responsible for coordination and oversight of the public works.

The Ministry of Finance and Economic Development (MOFED) oversees financial management of the programme and disburses cash resources to implementing federal ministries and to the regions based on the annual plan submitted by MOARD.

These federal implementation arrangements are replicated by regions and woredas (the Government's smallest administrative unit). In addition to programme implementation, regional and woreda bodies are responsible for ensuring sound multi-sectoral coordination of the public works. Public works planning and selection of PSNP beneficiaries occur within communities and kebeles.

Within this framework, non-governmental organisations (NGOs) and the World Food Programme (WFP) play important roles in implementation because of their experience delivering food aid and the institutional requirements of some donor agencies to channel resources through NGOs. In addition, NGOs and WFP provide technical assistance to the programme, and WFP supports the Government in procuring food stocks from abroad.

Donor agencies have pooled their financing - both cash and in-kind contributions - and formulated a unified stream of technical advice in support of a single programme led by Government. This has been led by the creation of a Donor Co-ordination Unit, expressly for the purpose of harmonising donor positions, providing technical support to the programme and managing all donor-to-donor and donorto- Government processes. The rights, obligations and coordination arrangements of the government-donor partnership for the PSNP are articulated in a Memorandum of Understanding (MoU).

Programme scale and coverage

In 2009, the PSNP supported 7.6 million people in 290 chronically food insecure woredas in eight of the country's 10 regions. This is equivalent to roughly 10% of the national population, covering over 40% of the country's woredas. The geographic coverage of the PSNP is shown in Map 1.

The 2009 annual budget was 2,136,734,460 birr in cash and 457,966.21 MT of cereals. This is equivalent to approximately $360 million - or about 1.2% of Ethiopia's GDP. In addition, the Government estimates that roughly $54 million in government staff time is devoted to the programme annually. There are at least 1,780 regular staff and Technical Assistants working on the PSNP full-time, with support from 14,295 Development Agents (DAs). The PSNP is integrated into the national budget system in Ethiopia.

Selection of beneficiaries

Chronically food insecure households residing in PSNP kebeles are eligible to participate in the programme. The PSNP Programme Implementation Manual (PIM) defines a chronically food insecure household as:

1. Households that have faced continuous food shortages (usually three months of food gap or more) in the last three years and received food assistance prior to the commencement of the PSNP.
2. Households that have suddenly become more vulnerable as a result of a severe loss of assets and are unable to support themselves (last 1-2 years).
3. Any household without family support and other means of social protection and support.

Based on these criteria, households are selected to participate in the PSNP through a community- based selection process. Once selected to participate in the PSNP, households are assigned to Public Works or Direct Support depending on the number of able-bodied members.

According to the PIM, households graduate from the PSNP when they are food sufficient, which is defined as "when a household is able to feed itself for 12 months a year, in the absence of programme support, as well as being able to withstand modest shocks." A household's food security status is assessed using a set of predetermined regional "asset-based benchmarks" that are tailored to local conditions. Data are collected by DAs and verified by kebele and woreda officials and vetted in community meetings.

Regardless of whether they are PSNP participants, any households that are unhappy with the outcome of the targeting and graduation processes are entitled to bring their grievances before the PSNP Kebele Appeal Committee (KAC).

Type, amount and timing of transfers

Transfers are provided to households on a monthly basis for six consecutive months. In 2009, the daily cash wage rate was 10 EBR and the food transfer was 3 kg of cereal. Each Public Works household member is entitled to receive a transfer based on five days of work at the prevailing cash or food wage rate. Households are provided transfers of cash, food, or a temporary mix of both resources. The mix of cash and food resources tends to be used in a way that addresses the seasonal rise in food prices leading up to the hungry season.

Public works are carried out from January to June, during the agricultural slack season, which is relatively dry. Transfers are delivered on a monthly basis.

Operating procedures for public works

Public works are identified and planned through a participatory process at community level. The Government's Community-Based Participatory Watershed Development Planning (CBPWDP) approach is the basis for developing a pipeline of sub-projects, many of which have a soil and water conservation focus. The woreda public works plan is integrated into woreda development plans to ensure linkages with other sectoral investments.

Tailoring the PSNP to specific groups

The PSNP includes a pastoral programme that addresses the different risks and vulnerabilities of pastoral livelihoods in the regions of Afar and Somali and pastoral areas of Oromia, and Southern Nations, Nationalities and Peoples Region (SNNPR). The programme uses public works and transfer payment mechanisms tailored to the needs of pastoralists. It is also designed to fit more organically with the institutional structures in these areas.

PSNP responds to the heavy workload of productive and reproductive labour carried by women in rural Ethiopia by allowing women to work fewer hours than men for the same pay and permitting them to switch to Direct Support when pregnant or breastfeeding.

PSNP risk financing mechanisms

The PSNP includes a contingency budget equivalent to 20% of the base programme cost and a risk financing facility designed to respond to transitory needs in chronically food insecure woredas. When households are unable to meet their immediate food needs because of shocks, these households are described as being transitory food insecure.

Of the contingency budget, 15% is held at regional level and 5% at woreda level. The woreda contingency fund is used to address unexpected needs of chronically food insecure households, such as a successful appeal to the KAC, and transitory food insecurity among PSNP and non-PSNP households. The regional contingency fund is used to address transitory food insecurity among PSNP and non-PSNP households in PSNP woredas.

The Government is currently operationalising the Risk Financing Mechanism to address transitory needs in PSNP woredas. This mechanism is designed to scale up the PSNP to cover transitory food insecure kebeles as needed, as well as to extend the duration of benefits for existing PSNP households, if needed.

Monitoring and evaluation

The PSNP key performance indicators are provided in Box 1. Given the flagship nature of the PSNP, much attention was, and continues to be, devoted to generating robust evaluations. A set of impact evaluations aims to measure changes for direct and indirect beneficiaries.

Design issues and trade-offs

In 2004, the Government proposed a caseload of 5 million chronically food insecure individuals. Donor agencies saw this figure as a minimum, with the actual number of chronically food insecure people probably closer to 7.9 million. The PSNP focused on those regions and woredas that had received food aid for the preceding three years or longer. Woreda administrators then selected the chronically food insecure kebeles, assigning the woreda's 'PSNP quota' from among these areas.

Within targeted woredas, chronic food insecurity was widespread but budget constraints required targeting households most in need. Within the food insecure woredas, household eligibility criteria were reinforced by a community- based selection process endorsed through community meetings. In addition, the wage rate for public works sub-projects was set low so that better-off households would be less tempted to seek employment through the PSNP.

The annual retargeting was designed to correct for inclusion and exclusion errors and thus responded to changes in the relative positions of households. A contingency budget of 20% was added to the programme to cover additional households that might become chronically food insecure during the course of the program and to respond to transitory needs among PSNP and non-PSNP households. In 2006, this was augmented by the Risk Financing Facility. The emergency response system would cover any transitory food insecurity beyond the available contingency and risk financing resources.

The emergency response system would also continue to cover food insecurity in non-PSNP woredas.

Targeting the PSNP

In 2005, the number of households selected to participate in the PSNP by communities and kebeles was almost always larger than the quotas assigned by the region and woreda. Recognising the scale of exclusion in programme areas, the Government and donors agreed that the October 2005 community-based targeting process would identify the actual number of chronically food insecure individuals. The Government and donors were therefore confident that the programme could handle a larger caseload, which was eventually (2009) set at 8.29 million people. The increase in beneficiary numbers eased the pressure on the targeting system significantly.

Evidence from 2008 shows that the PSNP is well targeted to the poorest households in PSNP woredas, which have significantly lower incomes, fewer assets and farm less land than non-beneficiaries (Table 2). A 2008 survey of local service delivery in Ethiopia reported that over 85% of respondents described the PSNP selection process as being fair.

Responding to transitory needs

In 2005 and 2006, the contingency budget was mainly used to cover additional chronically food insecure households and only evolved into a transitory response instrument in 2007. In 2007, the Government proposed that the regional contingency budgets cover transitory food insecure households in areas where the numbers identified by the emergency needs assessment were within this resource envelope. An emergency appeal would then be launched only for areas outside the PSNP.

Since the start of the PSNP, the relationship between the emergency response system and the PSNP has been far from clear. Extensive dialogue between the FSCB, DPPA and donors finally led to an agreement in mid-2008 that available regional contingency budgets would be used to respond to emergency needs arising from the failure of the small rains (belg) and food price inflation. The number of people benefiting from contingency resources from 2005 to 2008 is shown in Table 3. These figures suggest that the population covered by the regional contingency budget has grown in comparison with the regular PSNP programme.

Types of transfers

Implementation experience

Starting in 2004, food prices in Ethiopia began to rise, spiking in mid-2005 and again in mid- 2006. In both years, the Government and donors responded by accommodating the requests of woredas, which had planned to provide cash, to switch to food transfers. Given this experience with mid-year food price increases, many regions requested the Federal Government shift woredas from cash to food transfers in 2007. The programme accommodated this by allocating the first two transfers in cash rather than in food, which freed up food to be used later in the year in other woredas, and through a local purchase of food resources to augment donor pledges. Unlike previous years, few mid-year requests to change the cash-food split were accepted. The geographic distribution of cash and food transfers by woreda in early 2007 is shown in Map 2.

In 2008, the cash wage rate was increased for the first time from 6 to 8 birr. The trend in 2009 was towards greater requests for food transfers because of continuing increases in food prices. Given the resource mix available to the programme, coupled with budgetary constraints, the Government chose to allocate cash transfers for three months and then food transfers for three months to the vast majority of woredas. The cash wage rate was also further increased from 8 to 10 birr.

Since 2005, the overall trend has been increasing use of cash transfers in the PSNP (Figure 1) and a growing preference amongst beneficiaries for food only transfers (Figure 2). This move towards cash was, however, less than that initially predicted in the programme design and has largely stalled in recent years, due to global food and fuel price rises and parity issues between cash and food. In 2009, roughly 6.65 million people received food combined with cash transfers.

PSNP public works

While the PSNP design identified a menu of eligible investments, there was a tendency to focus on natural resource management projects.

One of the leading causes of previous failure of food for work programmes was the lack of appropriate community consultation to ensure the relevance and ownership of the assets created. In order to address this, the design of the PSNP gave the responsibility for the selection of public works sub-projects to communities. To promote this high level of community participation in the public works planning process, in coordination with Government DAs, the Community Food Security Task Forces (CFSTF) were mandated to mobilise communities to identify the public works that would be undertaken on an annual basis. In order to formalise this approach, DAs were provided training on the Community-Based Participatory Watershed Management Guidelines that were developed by Government in coordination with WFP.

Implementation experience

Since 2005, the PSNP has built a large number of public works sub-projects and evidence suggests that the performance of the public works component has improved over time. The 2008 Public Works Review found that quality had improved overall, with some continued technical deficiencies in roads and need for better operations and maintenance arrangements in water supply and small-scale irrigation projects. Sustainability of public works investments has been a challenge for programmes throughout Ethiopia. In general, sustainability ratings have been favourable, with the lowest ratings on roads, water and irrigation projects.

Despite these operational deficiencies, public works outcomes have generally been satisfactory with significant impact on environmental regeneration, increased access to water supply, expanded use of small-scale irrigation and improved access to farming training centres.

Direct support and tailoring the safety net

It was initially anticipated that roughly 80% of PSNP participants would receive aid through Public Works and 20% through Direct Support (unconditional transfers). While the number of Direct Support beneficiaries has increased, the Direct Support component as a percentage of programme participants remained constant (15%). This proportion varies by region and by year (Figure 3). In 2006, this proportion also varied among woredas, ranging from 3% in Enderta, Tigray, to 25% in Boricha, SNNPR.

The light work or community activities for Direct Support beneficiaries that were outlined in the PIM have not been implemented anywhere in the programme. Some woredas supported by NGOs have experimented with crèches or childcare centres. This highlights the lack of priority given to these activities, with woredas reporting that there had been no direction to include such activities in their PSNP plans.

Fieldwork observed progress in implementing the gender provisions of the PIM, although this was uneven across woredas and in elements of the PIM. Women and women's organisations tend to be well represented in PSNP decision-making structures at lower levels, however building alliances with the Women's Affairs Ministry at Federal and regional levels has proven to be more difficult. Although the PIM states that the participation of women in public works should be responsive to their regular work burden, in most cases men and women were found to do the same work and there seems to be no change in this over time.

In 2006, a separate process to design and implement a pilot programme that aimed to tailor the PSNP to pastoral livelihoods was launched. The pilot has demonstrated that a safety net is an effective means of supporting chronically food insecure pastoral households, particularly those who are engaged in agropastoralism or have dropped out of the traditional, mobile pastoral livelihood system. Experience from the pilot shows that the targeting mechanism needs to account for differences in social structures and social cohesion among different pastoral groups. With regards to public works, pastoral communities are easily mobilised to participate in the public works sub-projects that are relevant to their livelihoods.

Currently, the Government is eager to roll out the PSNP pastoral programme to all pastoral areas, particularly into new woredas in Somali.

Monitoring and evaluation (M&E)

The speed with which the programme was launched led to delays both in monitoring reports and the launch of the evaluation system. There was little information on programme implementation during the first months of the PSNP. In response, FSCB and donors were forced to establish rapidly an Information Centre to collect real-time information on cash and food transfers and grain market prices from a sample of 40% of programme woredas. These regular reports have proven to be an important source of information for programme managers, as the monitoring reports have continued to be very late and of limited use given the inconsistent quality and incomplete data.

The 2006 FSP M&E Plan reduced the amount of information collected, streamlined the reporting formats and modified reporting procedures. The focus was on generating basic information critical to programme management.

In 2007, the programme expanded the menu of independent systems reviews, augmenting the Roving Financial Audit with Roving Procurement Audit and Roving Appeal Audit. These audits were designed to generate information on the functioning of the programme safeguards.²

During the same period, consensus was reached on the need for a single logical framework. A series of joint government-donor workshops slowly created agreement on what the programme aimed to achieve and how this would be done. The result was a more harmonized approach to monitoring progress towards program objectives. This logical framework was used in each subsequent JRIS Mission to assess implementation and was the guiding document for the PSNP review in 2008.

As with other areas of routine monitoring, the PSNP was slow to develop a robust public works monitoring system as part of its management information system. By 2008, apart from intermittent monitoring through the public works reviews, there was still no database on public works projects and no tracking of public works standards, performance and effectiveness.

Key PSNP accomplishments

Main programme outputs

In 2009, the PSNP provided safety net support to almost 7.6 million rural Ethiopians. Importantly, this shows that large scale cash and/or food transfers are operationally and logistically possible in the resource-poor, lowincome rural settings in Ethiopia. The direct benefits of the programme are summarised in Table 4.

Key impacts on households

The 2008 Impact Evaluation found PSNP participation measurably improved household food security, as measured by changes in selfreported household food gap. This effect was strongest among those households that received regular, high value transfers. PSNP households exposed to drought had a 30% higher caloric acquisition growth than nonbeneficiaries.

From 2005/2006 to 2007/2008, negative coping strategies were more prevalent among non-PSNP households than they were among PSNP households.

Participation in PSNP public works increased growth in livestock holdings by 0.28 Tropical Livestock Units (TLU) over comparable non-PSNP households - equivalent to a difference of nearly three sheep.

PSNP beneficiaries have increased their use of social services. Of the 46.1% of PSNP beneficiary households reporting that they used health facilities more extensively in 2006 than the year prior, three quarters (76%) said this was because of the PSNP. In 2008, one quarter (26.7%) of households reported increased use of health facilities over 2007 and nearly half (47%) attributed this increase to the PSNP. In 2006, half (49.7%) of respondents stated that they kept their children in school longer than in the previous year, and 43% attributed this to the PSNP.

The 2008 Impact Assessment measured the combined impacts of the PSNP and the Other Food Security Programme (OFSP, the second pillar of the FSP focusing on providing credit services to PSNP households). Households that received high value, regular support from the PSNP and had access to the OFSP, experienced 0.81 months (or 25 days) greater food security from 2006 to 2008 than households receiving neither the PSNP nor OFSP. Similarly, livestock holdings of PSNP households with support from the OFSP increased by 0.334 TLUs, while the value of their livestock increased by 14.3% as compared to non-beneficiary households.

While OFSP has only a marginal impact on agricultural productivity when implemented alone, when combined with the PSNP the results showed a 38% increase in maize yields. This suggests that by allowing households to focus on long-term investments and providing more regular cash flow, the PSNP is a critical element of a strategy to effectively improve agricultural productivity.

Key impacts on communities

The bulk of investments were concentrated in soil and water conservation (SWC) and rural feeder roads, with selected investments in natural resource management and social services. In 2008, 92% of households indicated that their community had benefited from the construction of roads, while 88% reported benefiting from SWC on communal lands. Public works are increasingly perceived to benefit individual households as well. Public works are generally evaluated to be of a high technical standard.

Cost efficiency of the PSNP

The PSNP compares favourably with international experience on public works programmes, for its targeting, high wage intensity and a low administrative cost due to its use of existing government systems and the programme scale. Approximately 17.2% of total programme cost is dedicated to staff time, administrative costs and capacity building.

Graduation from the PSNP

Promoting graduation

To promote graduation, PSNP participants were to have access to the OFSP, which was financed through a Federal Government Specific Purpose Grant to regions and the donor-financed Food Security Project, amounting to roughly $100 million per year. Households were provided subsidised credit to rebuild their asset base (in the case of the Food Security Project which targets the poorest of the poor) or to purchase "household packages," which were various combinations of agricultural inputs sometimes based on a business plan developed with support from the extension service. The Government's Special Purpose Grant also financed investments in rural infrastructure, such as roads and water resource development, and the Resettlement Programme.

Recognising the complementary roles of the PSNP and OFSP to enable households to move out of food insecurity, starting in 2006 the Government specifically targeted OFSP household packages to PSNP participants. Government targets were set to achieve approximately 30% annual coverage of PSNP beneficiaries with the OFSP for three years. One of the main challenges to achieving this coordination was that the agricultural extension system was under-resourced and there were too few sufficiently skilled DAs. The governmentinitiated reform to upgrade this system was seen to be important for the success of the PSNP and OFSP, particularly the move to allocate three DAs to each kebele in the country and to ensure that posts are filled with people holding a diploma.

Initial experience with the OFSP found that delivery mechanisms were not always appropriate, which was reflected in low repayment rates and consequently low coverage. Specifically, there were no guidelines on how credit or revolving funds should be managed. When guidelines did exist, they were not always followed. By 2008, the Food Security Project had collected only 72% of loans that had fallen due and was working with the Government to transfer revolving funds to rural savings and credit cooperatives to ensure that they were properly managed.

The overall strategy to promote graduation has focused on households with available labour and land. Recently, the FSCB has expanded the scope of support to include nonfarm activities to better respond to the needs of young people, who generally have no land, in rural areas.

Determining graduation

In 2007, the Government initiated a process to set graduation criteria for the programme.

Thus, the Federal Government, regions and donors developed a set of objective asset-based benchmarks tailored to local conditions to measure a household's food security status. These benchmarks and an accompanying Graduation Guidance Note clarified that there were two levels of graduation: graduation from the PSNP upon obtaining food sufficiency and graduation from the FSP upon obtaining food security (Table 5).

While providing objective criteria against which to assess the assets of households, the overall system to identify households for graduation remained weak. Problems with communicating the benchmarks and graduation process resulted in widespread confusion. Additionally, the design of the system placed a heavy workload on DAs, who were regularly required to collect detailed household data.

Realistic expectations of graduation

Approximately 692,002 households (around 3.5 million people) received credit financed by OFSP and an additional 355,279 households received credit from the donor financed Food Security Project between 2002 and 2007. A number of independent studies have concluded that OFSP coverage was generally insufficient to meet the demand for loans among PSNP beneficiaries. While government reports suggest that access to a single household package should be sufficient to enable graduation, other evidence shows that the process towards graduation is more complex. PSNP beneficiaries indicated the need for two or three interventions per household to achieve food security.

There has been limited progress towards graduation to date. Between 2007 and 2009, around 280,000 individuals graduated from the PSNP. Although this is perhaps not insignificant given the adverse events of 2008, it falls well short of the national goal.

Looking Ahead

The next phase of the PSNP (2010-2014) is focused on strengthening implementation in all woredas. Measures include strengthening the monitoring system to ensure a regular flow of data to programme managers, adopting a more strategic approach to capacity building, reinforcing accountability and transparency measures at all levels and expanding these to the food management system and increasing the responsiveness of the program to transitory shocks. Greater attention to public works is anticipated to result in more sustainable public works sub-projects and enhanced programme impact within communities.

For the next phase of the PSNP, and more specifically the broader FSP, the Government has reformed the design of the OFSP in ways that should substantially strengthen it. These reforms have focused on the institutional arrangements for financial service delivery, including the development of multiple financial products tailored to the need and capacities of different types of households. Reforms also aim to strengthen the extension system and micro enterprise development programme to deliver market-led and demand-driven support to households.

Implementing this reformulated OFSP will likely result in higher levels of graduation from food insecurity within the next five years. While it is likely that the scale of the PSNP will reduce as households begin to graduate into food sufficiency, the notion of a 'floor' is central to this model - it suggests that a long-term social protection system in rural areas is important in order to both protect household assets and to create the type of agricultural production synergies currently witnessed between the PSNP and OFSP.

The next phase of the PSNP will consolidate and strengthen the RF mechanism to provide a predictable response to transitory food insecurity in PSNP woredas.

Conclusion

The PSNP clearly demonstrates the challenge of implementing a social protection programme in a low-income environment. Deficiencies in implementation resulting from limited human and physical capacity undermine the potential impact of the PSNP in many areas. Ensuring quality implementation in all programme areas remains a significant challenge for the future.

Nonetheless, Ethiopia's PSNP has demonstrated the value and potential of a transition from addressing food insecurity through the humanitarian response system to a system that is development-oriented.

Most importantly, PSNP has created, for the first time, a secure entitlement of households to a safety net from the Government. For the more than seven million people who receive PSNP transfers annually, this enables them to meet consumption needs, mitigate risks and avoid selling productive assets during times of crisis. As a result, there is evidence that livelihoods are stabilising and food insecurity is being reduced among beneficiary households. Similarly, it is increasingly apparent that the public works investments in soil and water conservation can result in significant improvements in the natural environment. This needs to be consolidated and built on.

For more information, contact Sarah Coll-Black, email: scollblack@worldbank.org, 
Matt Hobson, email: mhobson@worldbank.org and Wolter Soer, email: wsoer@worldbank.org

¹The coalition was established in 2003 to foster partnerships to take action to lessen the impact of droughts, improve livelihoods and ensure that communities have adequate food supplies. Funding support is from UNDP, OFDA and the World Bank.

²The term 'safeguards' refers to those mechanisms that aim to ensure that the programme resources are used for the purposes intended.