S P E C I A L   R E P O R T

FAO/WFP CROP AND FOOD SECURITY ASSESSMENT MISSION TO ETHIOPIA

(Phase 2)

Integrating the Crop and Food Supply and the Emergency Food Security Assessments

20 March 2008

• Mission Highlights
• 1. Introduction
• 2. 2007/08 Meher Production and the Overall Food Supply/Demand  Situation
• 3. Economic Situation and Price Trends
• 4.  Food Security and Emergency Needs at Household Level
• 5. Reflections on the 2007/2008 CFSAM process and recommendations for  future CFSAMs and associated Meher EFSAs
• Annex I,       Annex II,       Acronyms and Abbreviations

Mission Highlights

1. Introduction

FAO and WFP have carried out an unbroken series of annual CFSAMs in Ethiopia, at the request of the Government, since 1995¹. In the early years, the short assessment report, which focused primarily on production of cereals and pulses for the main Meher season, domestic supply/demand conditions and food aid needs, usually came out before the end of December of the assessment year. This was also the case in 2002, given the urgency of providing basic information during a year of crisis. In recent years, most of the CFSAM reports for Ethiopia have been issued well into January, or even in early February, i.e. at least 6 weeks after the conclusion of the in-country fieldwork and preliminary debriefings. There have been a number of reasons for this:

• First, Ethiopia CFSAM reports have become increasingly comprehensive and detailed, often including special sections or annexes accounting for the disparity between the two main national production datasets (from the Ministry of Agriculture and Rural Development, and from the Central Statistical Agency);

• Secondly, the Meher EFSA, an exercise led by the national Disaster Prevention and Preparedness Agency (DPPA), in which WFP, FEWSnet, and a few other stakeholders participate, has provided the main basis for the analysis of food and non-food assistance needs, and thus been used as the chief source of information for the WFP contribution to the overall CFSAM report. The Meher assessment process has also become more comprehensive and detailed over time and, because of the politically sensitive nature of its conclusions, the report has often been released only after careful high-level review and deliberation².

CFSAMs have been confronted with this issue not only in Ethiopia, but also in several other countries where significant needs assessment exercises take place independently of the FAO/WFP mission. This has been the case in southern Africa, for example, where national Vulnerability Analysis Committees (VACs) are responsible for needs assessments, often well after the CFSAM has undertaken its own supply/demand analysis, based on the status of the agricultural sector, just before harvest time. For such cases, FAO and WFP have proposed adopting a three-stage approach, consisting of:

1. a macroeconomic, crop production, food supply/demand analysis (CFSAM phase 1);

2. an assessment of emergency food and non-food assistance needs (done in southern Africa by the VACs) and, ideally, based in part on the results of the first stage; and

3. a synthesis of the preceding macro and micro-level analyses (CFSAM phase 2). ³

This, however, requires that the findings from each stage be available by the start of the subsequent stage, while the whole process needs to be completed fairly quickly in order to provide timely information for decision-making. This approach was proposed for Ethiopia in 2006 and it was agreed to use the 2007/08 Meher season as a test case. CFSAM phase 1, the crop and food supply assessment, was undertaken from 19 November to 12 December 2007, the field work for the Meher EFSA was conducted between 17 November and 17 December, and a joint FAO/WFP team travelled to Addis Ababa for CFSAM phase 2 from 21 January to 2 February 2008⁴.

Since the results of the Meher EFSAs had not yet been officially released, the team relied on preliminary information, including the summary reports of the Meher assessment teams presented to the Early Warning Working Group (EWWG) in December 2007, carried out an update of the food supply/demand and food security situation, and undertook a review of the CFSAM stage 1 and Meher assessment processes in order to make recommendations for future CFSAMs.

Three weeks after the mission had left Ethiopia, the DPPA released a document entitled Affected Population and Food Aid Requirement Summary by Woreda, dated 21 February, providing figures for the numbers of emergency-affected people in woredas that are not covered by the Productive Safety Net Programme (PSNP) and that it proposed should therefore be provided with assistance – “relief” – through DPPA mechanisms. Addenda have accordingly been placed in section 4 and Annex I of this report to reflect these ‘after-end-of-mission’ data and their implications.

2. 2007/08 Meher Production and the Overall Food Supply/Demand  Situation

As presented in the CFSAM phase-1 report⁵, the overall agricultural performance of the 2007 Meher season was judged to be better than the previous year’s, due to improved yields from a slightly increased cultivated area in both main production zones and marginal areas. Improved yields were considered to stem from well-distributed rainfall on crops, higher financial returns to grain growers prompting increased investment in inputs, timely farming operations, as well as a generally timely supply of fertilizers, improved seeds and credit⁶. Overall, the Mission estimated total 2007 meher cereal and pulse production at about 21.5 million tonnes, some 7 percent above the previous year’s post-harvest estimates. It has been the fourth consecutive good harvest, generally speaking. With a Belg production in July/August 2008 conservatively anticipated by the mission at 600 000 tonnes, total domestic availability of cereal and pulses for 2008 was estimated at some 22 million tonnes. Table 1 presents the national food balance sheet prepared by the CFSAM phase 1 team.

Conclusion: Domestic availability of cereal and pulses for 2008 is expected to more than meet aggregate domestic requirements. The balance between domestic supply and demand – see Table 1 – allows for a very substantial increase in estimated annual per capita consumption to 196 kg of cereals and pulses (in cereal equivalent terms),⁷ a substantial quantity of informal cross-border exports, and a large increase in stocks (from 376 000 tonnes opening stocks to 684 000 tonnes closing stocks).

Table 1: National food balance sheet for 2007-2008 marketing year (‘000 tonnes)

3. Economic Situation and Price Trends

3.1 Macro-economic situation

Since 2002/03, real Gross Domestic Product for Ethiopia has risen by 48 per cent, in real terms. This unprecedented economic growth has been strongly supported by generally good agricultural production, particularly as of 2004, since agriculture contributes about 45 percent of GDP, but it has also been driven by other components of broad based economic growth. These have included substantial public spending and investment, private investment from both domestic funds and growing inflows of outside remittances, as well as with more easily accessible sources of credit. Donor funding also rose from 3.7 Million Birr in 2005/06 to 7.6 Million Birr in 2006/07. At the same time, poverty-reducing expenditure (applying to such sectors as agriculture, food security, health, education and roads) grew about 18 percent in nominal terms, while defence spending was cut to a record low level of 2 percent of GDP by late 2007.

Table 2: General Consumer Price Index (CPI)

3.2 General trends in prices

Ethiopia’s general consumer price index has risen steadily since 2002 (when grain prices reflected the very bad growing season) and has reached double-digit levels since 2005, see Table 2. The trend has been strongly influenced by the food component of the consumers’ basket, the food component having risen much faster in the last few years than the non-food one, see Table 3.

3.3 Cereal prices

Since the end of 2004, prices of main cereals have followed a steady upward trend without any significant post-harvest reduction, and have remained above the average level of inflation throughout 2007 – see Table 3 and Figure 1.

Table 3: Movement of Food, Non-Food and Cereal Price Indices 2001 - 2007

Figure 1: Food, non-food, cereal and general consumer price indices 2000 to 2007

3.4 The causes and implications of rising food prices

Price trends in recent years and, in particular, the fact that basic grain prices have remained high ever after several good years, are due to a combination of economic factors influencing effective grain demand and supply as well as the rise in oil prices and the increasing supply of money. Continued good economic performance in 2006, fuelled by increasing pro-poor investments and private consumption, the injection of cash into the rural economy through the Productive Safety Net Programme (PSNP) and the recent increase in salaries and pensions of public civil servants have boosted the domestic demand for grains, livestock products⁸ and other basic foods. At the same time, despite the good harvests obtained, volumes of grains available on markets may not have increased as expected or, at least, their sales have been more evenly distributed throughout the year, instead of being concentrated at harvest time, as was previously the case. Better financial capacity of farmers to retain and stock grains due to access to micro-credit, increased local purchases by governmental food security institutions, agricultural cooperatives and relief agencies, together with expanded domestic and external trade flows, seem to be the main factors behind a market supply more evenly spread over time.

Table 4: Imported Cereal Food Aid (tonnes)

Finally, the fact that cereal food aid inflows, and food aid in general, have been decreasing over the last few years – see Table 4 – has contributed to a tight domestic supply: food aid inflows in 2008 are anticipated to follow the trend.

In a context of rising prices, it is sometimes suggested that the strategies and practices of traders (in particular spatial and temporal arbitrage) might be to blame for the price increases. According to careful observers of the national cereal markets⁹, however, there is strong evidence to the contrary. The marketing of cereals and other basic commodities has become less centralized, more dynamic and based on a greater number of primary and secondary markets of rising importance (including in particular the wholesale markets in Nazareth, Mekelle, Dire Dawa, and even Awassa). Ethiopian farmers have become increasingly market oriented and sophisticated, to the point that traders have been complaining that ‘farmers now exert stronger market power’. It is significant that in a recent interview with rural producers, farmers were heard to be requesting:

• accurate and timely market information,

• training on commodity grading, and

• increased access to warehousing facilities.

The question, under current conditions of favourable production conditions, relatively high prices, and supply response by farmers, is whether Ethiopia’s factor markets (especially for rural labour and agricultural inputs), and product markets (for all basic agricultural commodities) are fully up to the task of sustaining increased investment into an agricultural sector now yielding better returns.

Despite the considerable increase in nominal grain prices, neither wheat nor maize prices had reached import parity levels in January 2008, due to the very sharp price increase for these commodities on the world market. On the contrary, the very high current level of international wheat prices indicates that Ethiopian wheat prices may be below export parity, which explains the concern of the Government not only over the potential price impact of local purchases but also about potential exports of wheat to regional markets¹⁰. In fact, the Government is trying to resolve a set of complex equations. In theory, priority should be given to local purchases when there is an aggregate surplus in-country and, ideally, such purchases have to be undertaken in a number of small procurement actions from the surplus areas to minimize the impact on domestic prices. Given that the levels of grain prices in Ethiopia, in spite of their sharp increases, remain below those of most surrounding markets¹¹, there indeed is some scope for carefully-managed local purchases. On the contrary, spending a great deal more to import all of the emergency relief requirements for redistribution within the country would add to domestic supply, thereby creating an even greater disparity between domestic and over-the-border prices, and hence additional incentives for outflows of grain.

Conclusion: When choosing between local purchases or imports to meet requirements for emergency food distributions, there is policy choice to be made. Public resources can be spent either for the benefit of surplus producers and needy consumers (through local purchases) or for the benefit of needy consumers and those involved in informal cross-border trade (through commercial imports for all requirements).

4.  Food Security and Emergency Needs at Household Level

The annual Meher emergency food security assessment (EFSA) was undertaken between 17 November and 17 December 2007: 20 multi-agency teams visited 54 zones and 6 special woredas. The teams reported back to Early Warning Working Group (EWWG) meetings on 13 December (in relation to most of the cropping regions) and on 21 December (in relation to the pastoral regions and Gambella). At the time of the mission’s visit, no overall report had been issued. It was understood that the DPPA had compiled the zonal and regional figures and submitted its analysis to the Federal Government which was in the process of discussing the findings with regional governments.

4.1  Results of the December 2007 Meher EFSA

The assessment teams’ reports identified people in need of emergency assistance in all regions and a total of 49 zones. Most of these needs were in woredas, or “pockets” within woredas, that had suffered “weather adversities” such as localized floods, hailstorms or erratic rainfall. In Somali Region, however, the food security situation is described as being very poor in the majority of woredas, and could become critical in many, as a result of the combined effects of poor rains, localized events, and the disruption of markets due to insecurity and counter-insurgency measures.

Annex I provides a compilation, prepared by the mission, of the estimates for the numbers of people in need of assistance, by woreda, as reported by the assessment teams in December 2007, together with figures for the numbers of PSNP beneficiaries in the woredas concerned. The link with the coverage of the PSNP is important because the policy of the Government is to reduce the need for appeals for emergency food aid by using the PSNP to address chronic food insecurity and meet needs arising from localized, modest shocks – see the box below.

In fact, woredas fall into two main categories: those where the PSNP is operational, and those where it is not. In addition, within the non-PSNP areas, Somali region appears, not for the first time, to be a special case on account of insecurity in addition to poor rainfall, localized adverse events, the likelihood of a severe drought, and already rising levels of malnutrition.¹²

Different assessment methods were used in different areas

Existing household economy (HEA) baselines were used together with HEA livelihood impact analysis spreadsheets (LIAS) and woreda impact analysis spreadsheets (WIAS)¹⁵ for the analyses in Afar, SNNP, Somali and Tigray regions. An HEA baseline was prepared for Amhara shortly before the assessment and training sessions organized, but some of the teams and local officials were not sufficiently familiar with the HEA analysis to have confidence in the results and, in two zones, apparently gave precedence to ‘traditional’ analyses.¹⁶ For the remaining regions and special woredas, HEA baselines do not yet exist and the ‘traditional’ approach was used. Table 5 provides a summary of the methods used in different areas and the sources of technical support for the 2007 assessment.

The primary sources of information in all cases were officials and other key informants at the woreda level. Some teams, especially those using the HEA approach, also conducted interviews with market traders and a few – very few – community interviews.¹⁷

Table 5: Approaches used in the various regions and special woredas

Data were analysed and reported differently

Where the HEA approach was used, the spreadsheet analyses produced specific estimates of the impact of shocks on different livelihood groups and the numbers of people facing a ‘survival deficit’ (SD) or a ‘livelihood protection deficit’ (LPD) in each woreda. Some of the teams identified “emergency” cases that they judged could not be covered by the existing PSNP programme and contingency facility and/or numbers in “pockets” that they considered may not have been identified in the basic HEA analysis. But not all ‘HEA’ teams reported in this way and it seems unlikely that this was done in a standard fashion by the different teams.

Where the ‘traditional’ approach was used, discussions among the team and its interlocutors led directly to estimates of the numbers of people in need of emergency relief and the periods during which they would need assistance (the start date and duration). It is not clear from the reports whether, in PSNP woredas, the figures reported allow for some of the needs being covered through the PSNP contingency mechanism or not.

The variation in the presentation of the results in addition to the use of different methods in different areas makes it difficult to compile a reliable and coherent overall picture. Moreover, the mission found different figures for some zones¹⁸ and was not able to resolve all the differences, particularly for Amhara region and Sidama zone, and found only zonal – not woreda – level data for N Gonder and S Gonder. Some of the figures presented in Annex I are therefore open to question and may need to be corrected. However, the mission believes that Annex I provides an approximate, overall picture of what the assessment teams reported in December 2007 while also showing the different ways in which the assessment teams reported. The addendum (columns 12 to 16) also indicates the implications for the PSNP as well as DPPA.

Figure 2 shows the distribution of the woredas where emergency needs were identified and the overlap with the PSNP woredas.

Figure 2: Woredas where the Meher assessment teams found needy populations in Dec. 2007

Needs for supplementary food items

Table 6: Rations for DPPA and PSNP beneficiaries

The reports of some assessment teams, but apparently not all, assign categories (“moderate”, “severe”, etc.) to individual woredas and it is these categories that, in woredas not covered by the WFP-supported targeted supplementary feeding programme (TSF), determine whether blended food will be included in the DPPA rations for the emergency affected people or not – see Table 6. However, it is not clear what criteria have been used by the teams and it seems likely that there are considerable variations in the standards applied.

Many woredas have reported emergency conditions several times in the last few years

In reviewing the Meher assessment data, the Mission examined the status of woredas reporting “emergency” conditions by allocating them into PSNP and non-PSNP categories using the frequency of emergency conditions reported since 2005 (the launching of PSNP). The geographic distribution of the woredas frequently reporting emergency conditions and the overlap with the PSNP areas are shown in Figure 3; summary data are in the box below.

Figure 3: PSNP woredas and woredas that reported emergency conditions twice or more in the period 2005 to 2008

Most of the woredas with the highest frequency of emergency conditions are in Somali region and the planned expansion of the PSNP into that region may eventually alleviate the frequency of emergency conditions there.

The tendency for PSNP woredas to continue to report emergency conditions may indicate targeting issues within the woredas concerned.

4.2 Nutritional conditions

There is no systematic nutritional surveillance in Ethiopia. The periodic screening undertaken in 325 woredas by the UNICEF-assisted Emergency Outreach Strategy (EOS) programme should, in principle, provide indications of malnutrition rates and changes in those rates. But anomalies in some of the results undermine the reliability of these data.¹⁹

The Emergency Nutrition Coordination Unit (ENCU) in DPPA vets and compiles all the ad hoc nutritional surveys conducted by NGOs, but these are generally undertaken in response to reports of nutritional problems in specific areas and do not provide an overall picture of the nutritional situation or trends. However, these surveys frequently indicate particularly high rates of malnutrition for Somali and Afar regions, and in Sidama²⁰, and this has been the case in recent months. The relevance of rates reported from Somali and Afar has apparently been contested by some national authorities on the grounds that these nomadic groups are naturally tall and thin, but the evidence clearly indicates a deteriorating nutritional situation, with global malnutrition rates in Fik and Hamero, for example, already higher than measured at any time since 2003.²¹ The nutritional situation there is described as ‘critical’, taking account of aggravating factors such as drought, poor vaccination coverage and the insecurity that has been restricting commercial trade and food access.²²

The mission also heard some anecdotal accounts of deteriorating nutritional conditions in other areas. The situation clearly needs to be monitored carefully, not only in Somali region, but also in Borena, Afar and in other areas where severe emergency conditions have been reported.

4.3 The current situation, prospects, and needs for emergency assistance

Notwithstanding the overall good production described in the CFSAM phase-1 report (see section 2 of the present report), it is clear that the food security situation is serious in Somali Region and in some other areas, especially in the lowland areas of Borena, and that the situation in those areas could deteriorate further in the coming months. There are also needs in other areas.

Acute humanitarian needs in the Somali Region

The Somali Region DPPB Food Security Flash for December 2007, dated 6 Jan. 2008, described the result of poor rains and damage from locusts and estimated the number of people facing survival deficits at 745 000 with a further 842 000 facing livelihood deficits and also requiring emergency assistance. In addition, the mission saw and heard reports from staff of WFP and other agencies, who had visited various parts of Somali region in December and January, referring to water shortages, livestock stress and unusual livestock migrations as well as delays and difficulties in organizing food deliveries and distributions. It is understood that a 6-month emergency preparedness and response plan is being prepared by the regional government, to be shared with partners shortly. ²³

Based on information available, the mission believes that the disruption of markets due to insecurity and counter-insurgency operations constitutes a major threat to the food security of populations in large parts of the region, compounding the effects of water shortages, crop losses and poor rangeland conditions. The mission, therefore, learned with regret of the postponement, for lack of government clearance, of the proposed WFP-FAO mission to assess market conditions in the region that had been planned for the last week of January.

Poor and deteriorating conditions in Borena and neighbouring zones

The assessment team estimated there were 88 000 people in need of emergency assistance in Borena zone, and the mission heard reports of increasingly serious drought conditions and humanitarian concerns there. In the lowland part of that zone, insufficient October and December rainfall was said to have spurred a surge in disease outbreaks and livestock deaths while dwindling water supplies and human disease outbreaks were having a negative impact upon health, nutrition and wellbeing. Conditions and vulnerability levels were said to be similar in parts of Guji, Bale and East Haraghe zones.²⁴

The numbers requiring assistance in Borena are substantial in comparison with the PSNP beneficiary caseloads in the woredas concerned and the required level of assistance will require allocations from the regional contingency budget.

Pockets of need in other areas

There are undoubtedly pockets of need in other areas as a result of localized crop failures. However, the mission found it difficult to reconcile the widespread incidence of such pockets of food insecurity due to weather adversities reported by the Meher assessment teams, and the numbers of people affected, with the overall picture of good crop production for the same regions presented in the CFSAM phase-1 report (see, for example, the box below). Few additional and more recent data were available to the mission to determine whether any updating of the findings might be needed. However, the mission recently received indications that conditions are worsening some woredas in Gambella and E Gojam²⁵ (Amhara), and further similar developments may be anticipated as the Belg rains are already late for some of the Belg producing areas.

Aspects that could benefit from further examination

“Verification” missions are supposed to be organized by the regions to vet local requests for PSNP contingency resources in excess of the 5 percent of the current PSNP allocation that is at the discretion of the woreda. The mission was informed that such missions may also be organized to verify needs wherever the Meher EFSA reported significant numbers of people expected to need emergency assistance during the coming year. For the time being, there are no specific guidelines for such verification exercises but the mission was aware that a draft has been prepared (by DPPA and the EWWG) and is being reviewed and considered by the Government. It stands to reason that verification assessments should be at least as rigorous, objective and transparent as the original assessment, and should involve all the main stakeholders.²⁶

Some aspects that merit further examination in the context of verification missions and response programming are listed in the box below. They are important, amongst other things, to ensure reasonable equality of treatment for emergency-affected people in all areas regardless of whether they are in PSNP woredas or not, or whether their situation and needs were assessed using a traditional approach or an HEA-based approach.

4.4 Affected population and food aid requirements - Addendum (as of end-Feb. 2008)

The addendum to the table in Annex I shows in columns 12 to 15 the numbers of people needing relief assistance through DPPA mechanisms, and the numbers of people identified as needing close monitoring, reproduced from Affected Population and Food Aid Requirements by Woreda, DPPA 21 Feb. 2008. It also shows, in column 16, the numbers of people identified by the assessment teams as needing emergency assistance in woredas covered by the PSNP. These woredas do not appear in the DPPA listing: it is assumed that the PSNP is to cover all the emergency-related needs in these woredas. In line with the approach taken by DPPA for the non-PSNP woredas, the numbers of people to be covered through the PSNP correspond to those for “total emergency” needs in the case of “traditional” assessments and “survival deficits” only for “HEA-based” assessments. However, people facing “livelihood protection deficits” should also receive some level of assistance otherwise their future food security will be undermined.

There are some differences between the figures reported by the assessment teams and the provisions made in the DPPA requirements summary. It is assumed that many of these differences may arise from typographical, transcription or computational errors in one or other set of figures, but some may be more significant. The link between the two sets of figures is fairly clear in many cases. It is less evident for Amhara region²⁸ and Sidama zone, in particular, and the figures for some woredas in Somali region are significantly less than the even the survival deficit reported by the assessment teams (e.g. in Afder and Liben zones).

Table 7 provides a summary by region of the numbers of people requiring relief assistance through DPPA or close monitoring, reproduced from the DPPA document, and the numbers then remaining to be assisted through the PSNP in PNSP woredas. It also shows the numbers of woredas concerned. For PSNP woredas it also shows, in column 7, the number of people to be covered as a percentage of the normal PSNP caseload.

Table 7: Numbers of people and numbers of woredas requiring “relief” assistance through DPPA or close monitoring, and numbers remaining to be assisted through the PSNP

For the detailed, woreda breakdown, see Annex I.

The numbers requiring assistance are well within the 20 percent contingency budget for all regions except Afar. However, the 20 percent contingency provision within the PSNP relates to resources, not beneficiaries. If emergency beneficiaries need assistance for fewer months than normal PSNP beneficiaries, and if some of the people in need of emergency assistance are PSNP beneficiaries and therefore need only a small increase in the number of months for which they receive support, the percentage increase in the resources required will be less than the figures shown in column 7 (and in column 17 of Annex I). Further information is required to determine whether the requirements for Afar fall within the 20 percent contingency provision or an emergency appeal by DPPA is required. ²⁹

The figures in column 17 of Annex I show that, in almost all of the woredas concerned, the estimated numbers of people requiring emergency assistance are well above the 5 percent contingency provisions at the discretion of the woreda food security task forces. Notwithstanding adjustments for the number of months of assistance and the overlap with normal PSNP beneficiaries, factors referred tonne reductions, the lengthy, 9-step process of requests and approvals described in section 4.6.6 of the PSNP Programme Implementation Manual (Abridged version for Woreda use) will apparently be required to obtain resources from the regional contingency budget for most of the PSNP woredas.

Table 8 provides a summary by region of the food aid requirements for people to be covered through DPPA “relief” while Table 9 provides equivalent figures for people in PSNP woredas that are left to be covered through the PSNP. For more details of the basis of the figures, see the addendum in Annex I.

Table 8: Food aid requirements for people to be covered through DPPA “relief” (all quantities rounded to the nearest tonnes)

The estimated requirements for people to be covered through the PSNP (Table 9) are only “rough estimates” assuming that the emergency-affected people in PSNP woredas should benefit from the same levels of provision as those in non-PSNP woredas. The actual requirements would need to be calculated taking into account the numbers of months for which assistance is required in each woreda and, for blended food, the number of woredas outside the TSF where the situation is categorized as severe or very severe.

Table 9: Food aid requirements for people in PSNP woredas to be covered through the PSNP

(all quantities rounded to the nearest 10 tonnes)