OVERSEAS DEVELOPMENT ADMINISTRATION

EVALUATION REPORT Ev568

EVALUATION OF ODA'S RESPONSE TO THE 1991-1992

SOUTHERN AFRICAN DROUGHT

April 1995

Edward Clay, John 80rtol1, Sanjay Dhiri Anil Das Gupta Gonzalez and Carlo Pandolfi

VOLUME I

EVALUATION DEPARTMENT OVERSEAS DEVELOPMENT ADMINISTRATION 94 VICTORIA STREET LONDON, SW1E SJL

PREFACE

ACKNOWLEDGEMENTS

ABBREVIATIONS & ACRONYMS

SUMMARY

1.

2.

3.

4.

5.

6.

BACKGROUND

THE 1991/92 DROUGHT, ITS IMPACTS AND NATIONAL AND INTERNATIONAL RESPONSES

ODA'S RESPONSE

ASSESSMENT OF IMPACTS IN RELATION TO OBJECTIVES OF ODA'S RESPONSE

MANAGEMENT AND POLICY ISSUES IN THE RESPONSE

CONCLUSIONS

APPENDIX: TERMS OF REFERENCE

LIST OF TABLES

i

i i

i i i

1

13

15

23

29

41

55

65

Table 2.1 Index of Cereals Production in Southern African Region 16

Table 2.2 Southern Africa: Cereal Imports and Food Aid Requirements 20 for 1992/93

Table 3.1 ODA bilateral aid to drought-affected countries in 26 Southern Africa

Table 3.2 ODA bilateral response to the Southern African Drought 27 by channel of assistance

Table 3.3 ODA bilateral response to 1991/92 Southern African Drought 28 in 1991/92 by recipient country and sector

Table 5.1 Analysis of timeliness of ODA's response to the 1991/92 46 drought in three case study countries

Table 5.2 Analysis of timeliness of ODA's response to the 1991/92 46 drought in Lesotho, Mozambique, and Zimbabwe by sector

Table 6.1 ODA bilateral response to the Southern African 56 drought by recipient sector

Abbreviations & Acronyms

Annex 1

Annex 2

Annex 3

Annex 4

Annex 5

Mozambique Case Study

Zimbabwe Case Study

Lesotho Case Study

Terms of Reference

References

Volume II

i i i

1

37

77

97

101

Each year the Overseas Development Administration (ODA) commissions a number of ex post evaluation studies. The purpose of the ODA's evaluation rrogramme is to examine rigorously the implementation and impact of selected past projects and to draw out and highlight the lessons learned from them so that these can be applied to current and future projects.

The ODA's Evaluation Department is independent of ODA's spending divisions and reports directly to the ODA's Principal Finance Officer.

Evaluation teams consist of an appropriate blend of specialist skills and are normally made up of a mixture of in-house staff, who are fully conversant with ODA's procedures, and independent external consultants, who bring a fresh perspective to the subject-matter.

For this evaluation the team consisted of the following:

Edward Clay - Overseas Development Institute - Project Leader John Borton - Overseas Development Institute Sanjay Dhiri - Overseas Development Institute Anil Das Gupta Gonzalez -Carol Pandolfi - Consultant Civil Engineer.

The evaluation involved the following stages:

initial desk study of all relevant papers;

consultations with individuals and organisations concerned with the project, including a field visit to collect data and interview those involved;

preparation of a draft report which was circulated for comment to the individuals and organisations most closely concerned;

agreement with the evaluation team on the final report, which is published together with a summary sheet (EVSUM).

This process is designed to ensure the production of a high quality report which draws out all the lessons.

Johnny Morris Head, Evaluation Department

This study was prepared jointly by Edward Clay, ODI, acting as project leader, who also undertook a review of the overall response by the international community as well as ODA and field visits to Zimbabwe; John Borton, ODI, who reviewed ODA's food aid, and the response in Mozambique, as well as participating in a field visit to Zimbabwe; Anil Dasgupta Gonzalez, formerly SCF Country Director in Mozambique, who undertook the Mozambican field study; and Carlo Pandolfi, consultant civil engineer, who was responsible for the Lesotho field study and the review of emergency water projects in Zimbabwe. Mr Jeremy Shoham, London School of Hygiene and Tropical Medicine, advised on food and nutrition issues, Austen Davis, formerly employed by MSF-Maputo, provided prel iminary information on Mozambique, and Ms Pilar Guisasola assisted with beneficiary interviews in Mozambique. Sanjay Dhiri, ODI, was involved in file review, statistical analysis and the preparation of the main report. Catherine Hollis, Sue Horrill and many others provided secretarial and administrative support, and Margaret Cornell gave editorial assistance. The authors would like to acknowledge the many ways in which staff in ODA facilitated the study, as well as the information and assistance with field study arrangements in the three countries provided by the British High Commissions and Embassies. Officials of the governments of the three case study countri es co-operated fully wi th the evaluation. The staffs of NGOs and international agencies receiving support from ODA provided considerable relevant information. In addition, the help and advice of the staff of a number of organisations should also be acknowledged, including the USAID (FEWS) Project (Ha rare); WFP Country Offi ce, (Ha rare); Afri care; the National Consultative Forum on Drought (Johannesburg); SADC - Food Security Unit; SA DC Meteorological Information Centre; SA DC - FAO Regional Early Warning System (Harare).

i i

AGRITEX ATP BHC BoP BRCS CA CAFOD CADEC CARE CARITAS CC CCM CCPCCN

CENE CGIAR

COE CSAD CSFP CTE DANIDA DDF DEC DESA DHA DPCCN

DRIG DRSS DRTF DSM DWA ECFAD EDRMP EFF EMAD ESAP ESRT FAC FAO FAS FEWS FHI FPMU FRELIMO FSU GAA GIEWS GMB Gol GoM GoZ ICRC ICRISAT lOA IFRCS IMF

Ministry of Agriculture, Department of Extension (Zimbabwe) ODA's Aid and Trade Provision British High Commission Balance of Payments British Red Cross Society Christian Aid Catholic Fund for Overseas Development Christian Catholic Agency for Development An international confederation of NGOs An international confederation of Catholic NGOs Christian Care Christian Council of Mozambique Co-ordinating Committee for the Prevention and Control of Natural Calamities (Mozambique) National Executive Emergency Commission (Mozambique) World Bank's Consultative Group on International Agricultural Research Committee of Emergency Operations (Mozambique) ODA's Central and Southern African Department Child Supplementary Feeding Programme (Zimbabwe) Technical Committee for the Emergency (Mozambique) Danish International Development Agency District Development Fund (Zimbabwe) Disasters Emergency Committee (of British NGOs) Drought Emergency in Southern Africa Department of Humanitarian Affairs, (United Nations) Department for the Prevention and Control of Natural Calamities (Mozambique) Drought Relief Implementation Group (Gol) Directorate of Research & Scientific Services (Zimbabwe) Drought Relief Task Force (Gol) Dried skim-milk powder Department of Water Affairs (Gol) ODA's European Community and Food Aid Department World Bank's Emergency Drought Recovery and Mitigation Project Extended Financing Facility, IMF ODA's Emergency Aid Department Economic Structural Adjustment Programme (Zimbabwe) Evaluation Success Ratings Table Food Aid Convention Food and Agriculture Organisation of the United Nations Free alongside vessel Famine Early Warning System Food for the Hungry International Food Programming and Monitoring Unit of CARE Mozambique liberation Front Food Security Technical and Administrative Unit, SADC German Agro-Action Global Information and Early Warning System, FAO Grain Marketing Board (Zimbabwe) Government of lesotho Government of Mozambique Government of Zimbabwe International Committee of the Red Cross International Crops Research Institute for the Semi-Arid Tropics International Development Association International Federation of Red Cross and Red Crescent Societies International Monetary Fund

i i i

IRWS&S ITSH IWC JFS LCN LFDG LRC LSM MRCS MSF MSF-CIS NAC NANGO NCPCC NGO NOVIB

ODA 001 ORAP PEM RENAMO RSA SACU

SADC SCF SIDA SSFP TC UNHCR UNICEF UNILOG USAID VWS&S WA WASA WDC WFP WHO ZYMCA

Integrated Rural Water Supply and Sanitation (GaL) Internal transport, storage and handling International Wheat Council ODA's Joint Funding Scheme with NGOs Lesotho Counci 1 of NGOs Logistics and Food Distribution Group Lesotho Red Cross Life Sowing Ministries (NGO) Mozambique Red Cross Society Medecins sans Frontieres Medecins sans Frontieres Celula Inter-Sec~oes information system National Action Committee (Zimbabwe) National Association of NGOs (Zimbabwe) National Civil Protection Co-ordinating Committee (Zimbabwe) Non-governmental organisation The Netherlands Organisation for International Development Cooperation Overseas Development Administration Overseas Development Institute Organisation of Rural Associations for Progress (Zimbabwe) Protein-Energy Malnutrition Mozambique National Resistance Republic of South Africa Southern African Customs Union (of RSA, Botswana, Lesotho, Swaziland and Namibia) Southern Africa Development Conference (formerly SADCC) Save the Children Fund Swedish International Development Agency Schools Supplementary Feeding Programme (Zimbabwe) Technical Co-operation United Nations High Commission for Refugees United Nations Childrens Fund United Nations Logistics Task Force United States Agency for International Development Village Water Supply and Sanitation (Gol) Water Aid Water and Sewerage Authority (Gol) Water Development Consultants United Nations World Food Programme World Health Organisation Zimbabwe Young Men's Christian Association

iv

1. BACKGROUND

1. The drought which affected most of Southern Africa during 1991/92 resulted in the largest and possibly the most successful internationally co-ordinated food and relief effort of recent times. The greater part of the international response occurred during the fifteen months after March 1992. With the more favourable 1992/93 rains agriculture rapidly recovered, making further large food imports unnecessary (para 1.1).

2. ODA decided to assess its contribution to the overall response to the drought as part of a wider review of its experience in providing emergency relief. The Overseas Development Institute was asked to undertake this evaluation. It included: a review of ODA's general approach, including preparation and organisation, and the speed of response; a costing of the major elements and qualitative assessment of value for money; a review of the impacts of the major components; an assessment of the appropriateness of the assistance provided in terms of balance between financial aid, food aid, funding of NGOs, and technical co-operation. The study also looked at the co-ordination mechanisms involved and how they affected ODA (para 1.2).

3. The study consisted of a review of ODA files and interviews with the staff of departments involved in the response, followed by detailed case studies i nvol vi ng fi el dwork duri ng January and February 1994 in three affected countri es, Mozambique, Zimbabwe and Lesotho, selected to reflect the wide range of circumstances and the different elements of OOA's response. The findings of the study are set out more fully in the Main Report and Annexes 1 to 3 which cover the detailed case studies.

2. THE 1992 DROUGHT AND ITS IMPACTS

4. The ma in di rect impact of the drought was to reduce cereal production to below 40% of the late 1980s level, increasing import requirements by some 250% in 1992/93. Actual food imports of the severely affected SA DC countries and South Africa, including both concessional and commercial flows, increased by 7.2m tonnes or 235% (para 2.2.3).

5. The direct economic impacts of reduced food production were further aggravated by wi der impacts on agri cultura 1 producti on and export revenues, as well as non-agricultural impacts which were concentrated particularly in Zambia and Zimbabwe. Consequently, the affected countries were confronted with both internal and external financing gaps to cover massive increases in cereal imports. They had to take measures to facilitate agricultural recovery and to respond to water- and health-related effects of the drought, while counteracting the wider negative economic consequences of the crisis (para 2.2.5).

1

6. Thus the governments of the region and the international community were confronted with a disaster of unusual scale, severity and complexity. There were differentiated relief problems to be addressed in countries with varied capacities to mount large-scale operations. The drought posed problems of organising recovery, especially for the worst hit small-scale farmers. Water­supply and health problems were also potentially severe. Finally, economic impacts were highly differentiated. An important issue, therefore, in assessing ODA's response is the extent to which it was sensitive to differentiated consequences of the drought and ensured that its assistance complemented other actions (para 2.2.8).

3. THE RESPONSE BY GOVERNMENTS AND THE INTERNATIONAL COMMUNITY

7. The well-organised early warning systems in Southern Africa successfully drew attention to the impending crisis at an early stage, and then provided rapidly updated assessments of the crisis as it evolved (para 2.3.3).

8. By late February 1992 governments of affected countries and the international communi ty recogni sed the potenti all y immense scale of the regi ona 1 cri si sand began to organise appropriate responses, including regional logistics involving the Republic of South Africa, and a co-ordinated appeal for both food and non­food assistance. However, this process, which involved co-ordination by the UN Department of Humanitarian Affairs and SADC, was long drawn out, culminating in an international pledging conference in early June 1992 (paras 2.3.4-6).

9. Regional Logistics Response: an unprecedented level of imports required the organisation of supplies and finance and critical physical co-ordination of transport within the region. The combined requirements of RSA and the 9 drought­affected SADC countries were estimated in April 1992 at over Um tonnes for 1992/93 and, in the event, 10.5m tonnes of food arrived in countries of final destination during 1992/93. Especially impressive were the large increases in imports to landlocked Zambia, Zimbabwe and Malawi; and the 2.2m tonnes, including 1.3m tonnes in transit to other countries, shipped through Mozambique which was still affected by long-standing conflict (paras 2.3.12-15).

10. The success of the response resul ted from the combined efforts of the governments of the region and the international community to establish requi rements, make commitments of food and organi se the 1 ogi sti cs of its delivery. International co-ordination was less successful in assessing non-food relief requirements as these were limited to the appeals of United Nations agencies. A third, relatively successful, parallel process of co-ordinated international economic assistance through country donor consortia, including both the IMF and World Bank, involved assessing additional financial requirements. A number of contingent factors, including concerns about the risk of disruption to the processes of political and economic reform in the region, also contributed to the success of the international response to the drought (paras 2.3.16-17).

11. The financial cost during 1992/93 of the overall response to the drought was around US$4bn (£2.4bn) including actions by governments of affected countries.

2

The international food import and logistics operation as well as other rel ief measures funded by the international community amounted to some US$1.8bn (approx £l.lbn in 1992/93). ODA's own direct bilateral response eventually amounted to some £50m, wholly in grant form. This was some 5% of the international response during 1992/3 which itself included significant amounts of loan funding.

4. ODA'S RESPONSE

12. The role of ODA has to be considered in terms of its contribution to the overall success of the response at a regional level. The main elements of ODA's response strategy were determined as a result of its initial assessment of the drought situation at the end of February 1992 (paras 3.4-8). Two separate but complementary objectives were distinguished: to contribute directly to alleviating the suffering of affected peoples and to assist countries in coping with the severe economic effects of the drought. A phased approach was adopted, involving an initial response combining both humanitarian and economic assistance. ODA then reassessed the appropriateness of its overall response in the light of more detailed international assessments of the scale and distribution of the drought impacts (paras 3.9-11). The four basic elements of ODA's initial response on 4 March 1992 of approximately £10m involved:

• financial assistance, to meet balance-of-payments costs of additional drought-related imports in Zambia and Zimbabwe;

• bilateral food aid, targeted on countries where emergency aid was considered the most appropriate response, initially Mozambique and subsequently Malawi;

• technical assistance, particularly in support of the regi onal logistics effort;

• the use of this early, substantial UK response to encourage and facilitate European and wider donor responses to the crisis.

13. Subsequently, a block of emergency aid funding was earmarked to support technical assistance and the activities of British NGOs, and additional support for the programmes of international agencies including, in particular, the WFP and UNICEF (para 3.12).

14. The phased response involved three tranches of funding and total commitments of £56m, including emergency assistance to Mozambique that was only in part drought-related (para 3.13).

15. Actual expenditure, excluding some assistance to Mozambique which was only partially related to the drought, totalled some £49.2m. In terms of regional resource flows, total bilateral assistance to the severely drought-affected countri es of the regi on increased by £30. 5m or 25% duri ng 1992/93, the maj or areas of increase in expenditure being programme aid to Zambia and Zimbabwe, disaster relief and food aid. Some 43% of ODA's assistance was channelled

3

directly to governments of the drought-affected countries, 36% indirectly through British-based NGOs and their partners, and 17% via international agencies (paras 3.16-21).

16. The sectoral composition reflects priorities determined by ODA in its initial response, and reactions to proposals from governments, NGOs and international agencies. The largest single element of expenditure, balance-of­payments support (41%), reflected ODA's assessment of the severe economic impacts in Zambia and Zimbabwe. Food aid for direct distribution, (35%) was the preferred response in Malawi and Mozambique. These decisions reflect at least implicit judgments about the capacities of national institutions in drought­affected countries. The other elements of the response (25%) to a large extent reflect the balance of proposals put directly to ODA by NGOs and international organisations, or channelled through the Posts (para 3.24).

17. The important issue for assessment is therefore the relative effectiveness of those allocations which reflected ODA's choice of priority, the provision of balance-of-payments support and the country targeting of food aid; and those activities where it was reacting to NGO and international agency requests (paras 3.25-30).

5. ASSESSMENT OF ODA'S RESPONSE WITH SPECIAL REFERENCE TO THE THREE COUNTRY CASE STUDIES

18. The response of the affected countries and the international community to the 1991/92 drought is widely regarded as a success story, and this evaluation broadly endorses that conclusion (see below paragraphs 36-40). The international community more generally and ODA in particular have succeeded in incorporating into their emergency practice many lessons learnt since the previous regional crisis in Africa in 1984/85 (para 6.2).

19. Nevertheless, a closer scrutiny indicates areas of less satisfactory performance. There is need for improvement in the way ODA, NGOs and the international community prepare for and respond to drought shocks given the likelihood of another major drought-related crisis in Africa within the next few years (para 6.3).

i. ODA's General Approach

20. ODA performed well in terms of quickly identifying the worst affected countries, establishing clear objectives and devising its overall response strategy at an early stage using the full range of aid instruments at its disposal. This was because CSAD, in consultation with the officials in post, effectively monitored the situation in the major countries of the region and ECFAD maintained a watching brief on food security and food aid (para 6.4).

21. The principal weaknesses lay in terms of the way ODA organised its response at a country level, where strategies were less clearly developed (para 6.5). In particular:

4

• it did not have a specific strategy for targeting its food aid in Mozambique and consequently the NGOs, through which virtually all its assistance was channelled, determined the geographiral distribution. In a difficult operating environment most of these agencies preferred to direct additional drought-related resources to their existing relief programmes whi ch were not in those areas most severely affected by the drought. It is not known whether, as a result, there were unmet needs in the latter areas.

• it failed to question rigorously the content of many proposals submitted by agencies, some of which were uninformative and poorly prepared;

• in-country monitoring capacity was insufficient to ensure a necessary minimum number of monitoring visits to the project sites and follow-up on agency reporting. In particular, most grants made to UK-based organisations, sometimes for partner agencies in-country, were not satisfactorily monitored (see below).

ii. Speed of Response

22. ODA initially moved quickly. Its first tranche of support on 4 March 1992 was made only a month after Zambia declared an emergency and before the Government of Zimbabwe had made its official declaration and request for international support (paras 6.6-6.8).

In terms of specific components:

• balance-oJ-payments support was provided speedily and, despite being a new form of emergency response, smoothly;

• the bulk (90%) of the Jood aid provided to Mozambique during 1992/93 was di stri buted by the end of 1992, before the start of the normal 'hungry season';

• the performance of non-food assistance including that in the agricultural recovery, emergency water programmes and heal th sectors, was 1 ess impressive. The major part of the water and health assistance failed to yield benefits to the target population(s) until after the 1992-93 rains had begun and the crisis had passed. Many of the seeds and tools provided as part of agricultural rehabilitation activities did not reach farmers until well into the planting season.

23. As some NGOs were slow in submitting proposals for non-food aid assistance, their actions often started comparatively late. The proposals of some NGOs, submitted on behalf of partner organisations in the affected countries, were of low quality, which caused further delay. ODA's performance in terms of the time taken to process and approve the proposals was satisfactory in most cases, especially taking into account the varying efficiency of those seeking funding and implementing activities.

5

iii. Value of the Major Elements and Value for Money

24. Overall, ODA appears to have obtained value for money (para 6.9). In terms of specific components:

• the majority of ODA food aid was distributed free, but there was also some food-for-work and food for subsidised sale by NGOs. Local purchases of food for relief, where direct comparisons could be made, compared favourably with costs of food aid imports by other donors;

• as agricultural rehabilitation projects were implemented relatively late, many agencies Ilad to pay a premium which could perhaps have been avoided had purchases been arranged earlier;

• in undertaking truck purchases, some agencies opted for apparently better deals available from South African suppliers but some of the models so procured may have been less appropriate;

• the procurement and shipping of drugs and other medical supplies by UNICEF appears, overall, to have been cost-effective.

But the issue of cost-effectiveness cannot be considered separately from that of timeliness.

iv. Impact of the Major Components

25. There were difficulties in determining the impacts of most ODA-supported actions in the three case study countries and more generally (para 6.10). Impacts can only be inferred indirectly, by the way in which affected groups, regions and countries recovered, without evidence of intensified problems of economic distress, eg migration and increases in morbidity and mortal ity. A further problem is that the interventions supported by ODA represented only part of the overall drought response, even at the district or provincial level.

26. The other reason for the difficulty in determining impacts is that very few of the operational agencies involved attempted to measure impact (para 6.11). Given the scale of the assistance provided and its importance to the intended beneficiaries, agencies should have striven to assess the impact and effectiveness of their relief actions more frequently, at least on a selective basis.

v. Assessment of the Major Benefits Delivered in Relation to Costs

27. Inevitably, given the lack of information on impacts, cost-benefit assessment is, at best, impressionistic (paras 6.13-15).

28. The benefi ts of financial aid for balance-af-payments support are inherently difficult to assess. Grants to Zimbabwe and Zambia, however, were extremely well-timed and fast acting. This assistance contributed to the success of the

6

governments' efforts in avoiding the breakdown of the normal food marketing system and the serious economic and pol itical consequences that would have followed. The additional assistance also decreased the governments' financing requirement for drought response, thereby reducing the drought's negative impacts on medium-term economic reform strategies.

29. Although the impact of food oid in Mozombique could not be assessed, the bulk of distributions were completed before the start of the next harvest.

30. The lack of information on the results of most seeds distribution programmes makes it impossible to assess whether the benefits outweighed the costs.

31. The late start-up and slow implementation of most of the emergency water programmes substantially reduced the value of their benefits during the drought. Even so, where the boreholes and other supply improvements were late, they have probably been of long-term benefit to the population. It is likely, however, that such longer-term benefits could have been achieved more cost-effectively had the activities been undertaken in a carefully planned way as part of a normal development programme (for example in Lesotho) rather than as an emergency intervention. The combination of food-for-work, as a relief measure, with construction of dams to improve village-level water supply (Christian Care, Zimbabwe) is an example of an activity which, by successfully combining immediate relief with longer-term goals, will command considerable local support in a crisis.

32. The health sector was not identified as a priority area for ODA support in its initial assessment, but, given the outbreaks of drought-related diseases, preventative measures should perhaps have been given a higher priority. The lack of integration between emergency water and health measures is also a less than satisfactory aspect of the drought response.

vi. Appropriateness of the Assistance Provided, Including the Balance Between Financial Aid, Food Aid, Funding of NGOs and Technical Co-operation Input

33. The innovative use of financial aid as BoP support in an emergency situation was a very positive action. Such a response is especially appropriate where the recipient government has the capacity and commitment to organise an effective national disaster response but where financing the response would involve economically damaging foreign-exchange costs. By assisting in this way, ODA also probably helped to galvanise international support for Zambia and Zimbabwe, in particular, to limit the negative impacts on their medium-term economic strategies (para 6.18).

34. Food aid was correctly targeted on Mozambique, where direct relief to victims of the drought, whose needs had been increased by the war, had the highest priority. However, the bulk of ODA food aid was distributed in areas other than those most affected by the drought and concentrated on the provision of cereals. There is evidence in aggregate that Mozambique received a higher proportion of its cereals needs than was the case for oils and legumes. Although

7

there is no quantitative evidence that unbalanced rations resulted in nutritional problems, some of ODA's NGOs partners expressed concern at ODA's concentration on the provision of cereals (paras 6.19-20).

35. A strength of ODA's emergency funding is its considerable flexibility and simple requirements for disbursement. This was generally noted as resulting in rapid utilisation. Such flexibility, however, implies a need for further reporti ng as part of ODA' s accountabi 1 ity for the use of publ i c funds (para 6.21) .

vii. Co-ordination Mechanisms Available, Their Effectiveness, and How These Affected ODA's Actions

36. The international and regional early warning systems functioned well in providing early information on the drought. The varied responses, however, of the governments of affected countries, and also of donors, indicate that the co­ordination of actions on receipt of such information is still far from satisfactory (para 6.22).

37. Food aid donors' extempore co-ordination was successful in initiating an adequate international response and support for the regional logistics operation, which ODA funded in a small way but also strongly supported in policy terms. It was an outstanding example of international co-operation (para 6.23).

38. The DESA process, which ODA also supported in a small way, was a less successful but still positive aspect of the international response. The two principal weaknesses of the DESA process were, fi rst, that it was an over­ambitious attempt at regional co-ordination which only delayed action, and, second, that it failed to integrate bilateral non-food aid channelled to governments and NGOs within the process (para 6.24).

39. The 'broker role' of some UK-based NGOs in seeking funding for their local partners from a number of donor sources, and thei r apparent 1 ack of co­ordination, meant that it was extremely difficult for ODA to know whether it was funding activities with the highest priority. Furthermore, it was sometimes difficult for officials in Post to ensure that there was no duplication (para 6.25).

40. NGOs, in retrospect, themsel ves emphasi sed the need for stronger co­ordination of their mutual efforts in organising drought relief (para 6.27).

viii. The Effectiveness of Monitoring and Accountability Arrangements

41. The country case studies consistently highlight weaknesses in monitoring and reporting on the part of a high proportion of the agencies operationally involved in the drought response. Where, however, the Post had an explicit role in monitoring grants, a high standard of reporting was noted (para 6.28).

8

42. Weaknesses in NGO monitoring and reporting, as well as in proposals, are considered to be part of the wider issue of transparency in the use of and accountability for public funds. International NGOs and consortia of NGOs, as well as international agencies, are drawing upon funding from many governments and from the general public across the Western world. Although the scope for a single donor such as ODA to establish priorities and to get agencies to improve proposals and monitoring is constrained, this obstacle should not be allowed to block more vigorous efforts in that direction (para 6.29).

ix. Success Ratings and Overall Success

43. The evaluators considered the criteria in Evaluation Department's Success Rating Table for summarising the assessment of ODA's drought response. Performance on technical success, cost-effectiveness and conditions on financial reporting was variable; support to economic liberalisation was successful; and the international effort was highly successful in protecting the reform process. While poverty impact was apparently successful there was insufficient information to assess the success for vulnerable regions and target groups and similarly for the health impact, impact on women and social impact (para 6.34).

44. The evaluators conclude that a set of performance criteria appropriate for assessment of a large complex emergency programme must differ from one devised for assessment of development projects. The assessment of disaster impacts was satisfactory, but the quality of proposals was extremely variable. An overall assessment of timeliness of a range of activities is impractical, but areas of variable performance included formulating and submitting requests. In terms of sectors, food distributions were relatively timely, seeds and tool distributions were variable and emergency water supply measures had little impact in relation to drought-created needs. ODA's own performance was satisfactory in terms of the time taken to review requests, and good in the provision of BOP support, but assessment of NGO proposals was variable. Sensitivity to changing conditions was a positive aspect of performance (para 6.35).

45. Finally, it was not possible to make an overall assessment of success. To do so it would have been necessary for higher priority to have been accorded to monitoring and to impact assessment during implementation (para 6.36).

LESSONS

46. In responding to a large scale regional emergency with a range of impacts, use of all available aid instruments increases the likelihood of a more effective response.

47. The provision of balance of payments support in grant form to a low income disaster-affected country is an appropriate response where there is:

• agreement between donor and recipient government on overall aid and public expenditure priorities and the reporting requirements are satisfied by arrangements already in place. It is more appropriate where the government

9

itself makes the major response, and where there is a related resource gap that would not be filled by existing international compensatory arrangements. Grant finance is of course particularly welcome to poorer countries in these circumstances.

.. Directly assisting a government's own import programme should have priority where this complements and underpins relief efforts by local and international NGOs. In Zimbabwe and Zambia NGOs were only able to launch relief operations in a timely way and to sustain these activities because they could draw upon marketing channels that were sustained by the government's own import programme.

.. A preference for financial assistance might also be contingent on the pragmatiC judgement that commercial imports are likely to provide a speedier response than conventional food aid.

48. For the provision of food aid support:

.. Initial targeting by geographical area and vulnerable populations are likely to be the critically important decisions.

• As well as responding to requests from familiar partners, the possibilities of providing food aid through less familiar partners should be considered. Choices here depend critically on co-ordination with other donors.

• Regarding sources of commodities, local purchases can provide flexibility and more timely responses. The appropriateness, however, of the commodities in question, as well as assessment of the local marketing implications, should be part of the assessment process prior to the making of a grant.

.. That a significant number of agencies more or less automatically turned to ODA to fund vehi cl e purchases or transport costs refl ects on ODA' s flexibility. That such activities can, when taken together, account for significant expenditures, suggests the need for a co-ordinated technical co-operation input into future large relief programme to ensure that decisions are made on an appropriate and cost-effective basis, and that equipment is effectively utilised.

49. If improving quantity and assuring potability of water supplies during a drought are priorities, then there are lessons to be learned from 1992/93 in terms of appropriate mitigation measures. Some other donor projects concentrated on rapidly implementable measures to improve supplies, such as well deepening and repairs, in order to have more impact during the drought. Slow start-up suggests that effective emergency water measures would benefit from more effective disaster preparedness. Institutional arrangements also make a considerable difference to effective response: the use of the NGO, Water Aid, as the implementing agency in a government-related programme, achieved more rapid implementation, some sensitivity to changing priorities, and more effective

10

monitoring than in other projects. Emergency measures to increase water supplies must be properly integrated with health considerations. There is also the difficult decision-making problem under conditions of uncertainty during a drought, over the balance to be struck between measures to improve water supply almost immediately and measures that would have considerably greater benefit in the long-term or if the drought were to continue.

50. There appear to be three 1 essons for future agri cultural rehabil itation, especially as operated by NGOs. First, such programmes and proposals should be prepared alongside those for relief measures because of long lead times and the likely poor performance of late distributions. Second, suitability and quality of seed are likely to be decisive considerations and NGOs may need technical support. Third, performance monitoring and impact assessment need to have higher pri ority.

51. The drought CrlS1S had a potentially serious health aspect, which was worsened by the war in Mozambique and Angola and the movement of people to centres with, initially, poor water, sanitary and health facilities. Some aspect of preparedness is desirable, once a crisis beyond even national dimensions is recognised. Donors such as ODA therefore have a useful role in providing additional foreign exchange to meet import requirements, particularly those of essential drugs related to disorders associated with contaminated water supplies and nutritional stress. The relationship between nutrition, water supply and health actions needs to be recognised, and in this regard monitoring has an important role.

52. Where, as in the case of a drought, agencies are not responding to immediate, life-threatening problems (for example, in organising food distributions over several months; providing seeds for agricultural rehabilitation; or assuring potable water supplies) there remains considerable scope for improving the quality of proposals and for strengthening monitoring, impact assessment and reporting.

53. Early contingent response by key donors is critical in ensuring overall success. International coordination during implementation, in particular at the regional level, is necessary for success in region-wide emergencies.

1 1

1.1 The drought which affected Southern Africa during 1991/92 resulted in the largest and possibly most successful internationally co-ordinated food and relief effort of recent times. The scale of food imports alone, both commercial and aid-assisted, represented the largest action of its kind since the Indian drought of 1966-67 and the aversion of famine in Europe in 1945-46. The greater part of the international response occurred during the fifteen months after March 1992. With the more favourable 1992/93 rains over most of the region, agriculture rapidly recovered, making further large drought-related imports unnecessary.

1.2 The UK Overseas Development Administration (ODA) decided to make an assessment of its contribution to the overall response. This in turn was seen as part of a wider review of ODA's experience in providing the emergency relief that had become an increasingly important part of its overall actions during the previous five years. The Overseas Development Institute (ODr) was asked to undertake thi s assessment whi ch represented a follow-up study to the three previous assessments of ODA's response to major natural disasters'. The evaluation was begun in November 1993 with fieldwork visits in January-February 1994.

1.3 The Terms of Reference are set out in the Appendix to this report. The method followed was broadly that of the three previous evaluations of ODA emergency responses. First, members of the team reviewed ODA files and interviewed the staff of departments involved in the response, namely the Central and Southern African Department (CSAD), the European Community and Food Aid Department (ECFAD)' and the Emergency Aid Department (EMAD). The second phase involved detailed case studies of three affected countries chosen to refl ect the wi de range of ci rcumstances and the di fferent el ements of ODA' s response: Mozambique, Zimbabwe and Lesotho.

1.4 In assessing ODA's contribution as part of the overall response to the drought, the approach adopted was to look at ODA's response as it related to ODA's own objectives. The first objective was to contribute directly to the alleviation of human suffering through the use of food aid and emergency grants. The second objective was to counteract the severe economic impact on countries in the region through bilateral aid funds. ODA was also concerned that its response shoul d be made in such a way as to promote an adequate and timely international response. Finally, ODA was concerned that its contribution should complement other international assistance in circumstances where it would be providing only a relatively small component of the total response. ODA's own direct bilateral response eventually amounted to some £50m in untied grant aid, 5% of the international response during 1992/93. 3 The issues for closer investigation in this evaluation were selected to focus on the way ODA's response fitted into the wider one. The evaluation also considers, therefore, the ways in which governments, multilateral organisations and NGOs sought ODA's support and attempted to integrate that assistance into their responses to the drought.

1.5 Agencies in receipt of ODA support did not, with few exceptions, accord a high priority to monitoring the impacts of actions in any detail. The main exception was the massive food import and logistics aspects of the response, of which there have been a number of assessments of operational efficiency.

These were the Ahican food crisis of 1984·86 (Borton et al, 1988); the Bangladesh floods in 1988 (Borton et ai, 1989); and the Bangladesh cyclone in 1991 (Jones el al, 1993).

Now renamed the European Union Department.

£5001 represented approximately US$85m at the time of disbursement. The financial cost during 1992/93 of the overall response to the drought has been estimated at approximately US$4bn including actions by governments of affected countries. The international food import and logistics operation, as well as other relief measures funded by the international community, amounted to approximately US$1.Bbn. This is an 001 estimate based on UN and IMF documentation.

13

Background Southern African Drought

Therefore, taking into account both the 1 imitations of the evidence and the attention given to that one aspect of the direct response, this evaluation has concentrated on providing a broad assessment of the whole of ODA's response, complemented by a small number of more in-depth investigations of selected grants, especially those concerned with emergency rural water supply. In examining the efficiency of the response, the evaluation has also closely examined the ways in which 'partner' organisations, especially NGOs, sought to access and integrate ODA's financial and food assistance into the drought emergency activities.

1.6 The drought and its impacts are described in Chapter 2 along with the response of governments and the international community. In Chapter 3, there is an overview of ODA's response. Chapter 4 covers the team's assessment of ODA's response, summa ri sing ev i dence of impacts of the ass is tance provi ded and focus i ng on balance-of-payments support for financing additional imports; food aid, distribution and logistics; emergency water programmes; agricultural rehabilitation, especially seed distribution; and health sector responses. Management and policy issues are considered in Chapter 5, including: organisation within ODA and the posts; ODA's strategy; the efficiency and timeliness of the response; its cost-effectiveness; and the effectiveness of the co-ordination arrangements and of the monitoring and reporting arrangements. Conclusions are summarised in Chapter 6. Lessons are set out in the report's Summary. The three case studies of ODA's response in drought-affected countries, Mozambique, Zimbabwe and Lesotho, are contained in Annexes 1-3.

1.7 There is an issue about which countries were affected by the drought, and so should be included in an assessment either of the overall international response or of ODA's bilateral actions. SADC, as a regional co-operation organisation, automatically includes assessments of the food situation, including food aid requirements, of all its member countries in early warning bulletins and regional appeals. Consequently, both Angola, where food production apparently rose during 1992/93, and Tanzania, which was little affected by the drought, were included in the regional appeals, whereas RSA was not expl icitly taken into account in assessing the regional food situation and was excluded from the appeals. For the purposes of this evaluation, the analysis is confined to the eight SADC countries which were both severely affected by what turned out to be 'the drought of the century' and in receipt of ODA bilateral assistance: namely Botswana, Lesotho, Malawi, Mozambique, Namibia, Swaziland, Zambia and Zimbabwe.

14

2.1 DROUGHT: A SLOW-ONSET DISASTER

2.1.1 Drought is widely characterised as a slow-onset disaster, in contrast to sudden-onset events such as earthquakes or cyclones. The slow-onset character of a drought has two aspects. First, it impacts directly on agriculture or other economic activity and human wellbeing, through deterioration in water supply and quality, typically intensifying over several months or even years. Second, the precise extent and intensity of these effects of rainfall variability are uncertain as the crisis evolves, only becoming fully apparent with the benefit of hindsight, once the drought has broken. Assessment of the appropriateness and efficiency of responses to drought needs to be based, therefore, at least on basic familiarity with the specific crisis and its characteristics.

2.2 AN OVERVIEW OF IMPACTS

2.2.1 The summer rainfall region of Southern Africa includes the greater part of RSA and the SADC group of countries. Drought as a meteorological or agricultural phenomenon is a quite frequent occurrence somewhere in the region. It occurs infrequently in some parts of the region and only occasionally, as in 1913 and 1991/92, regionwide. 4 Summer rainfall is concentrated between November and late March. The rains are 1 inked to the position of the Inter-tropical Convergence Zone, whi ch is pushed northwards away from the Southern Afri can region, preventing any significant rainfall for extended periods. The 1991/92 drought affected most of RSA and all the SADC countries except most of Angola and Tanzania and the northern part of Mozambique. It has been described as the worst drought in the region in living memory.

2.2.2 The risks of a regional crisis were high in 1991 because rainfall had already been below average in 1990-91 and food production had declined sharply in marketing year 1991-92 (Table 2.1). Poor early rains in October-November 1991, which gave rise to initial concern amongst agro-meteorologists, were the first definite indication of a potentially major crisis. However, the regional impact and intensity of the drought were by no means clear and, until the end of 1991, reports produced by SA DC and the FhO's Global Informdcion and Early Warning System were ambivalent. But, with an almost unprecedented dry spell in January­February 1992, such rains as eventually arrived were too late (and insufficient) to save most of the standing crop of maize throughout Zimbabwe, southern Zambia, the central and southern areas of Mozambique, all Botswana and Lesotho and large areas of RSA.

2.2.3 The main direct impact of the drought was on food production, thereby increasing food import requirements and threatening widespread rural distress and possible famine conditions. Maize production and even the more drought-resistant cereals such as sorghum and millet were devastated on a region-wide basis. The severely affected economies suffered a dramatic decline in cereal production during 1992/3 to only 38% of the late 1980s level. In regional terms, the most severe impacts were on the largest producers, RSA the main grain exporter of the region, which produced less than 40% of its 'normal' crop, and Zimbabwe, often an exporter, where national production was less than a quarter of the annual average during 1986/87-90/91 (Table 2.1).

The distinction between drought as a meteorologicaVhydrological/agricultural and as a social phenomenon raises complex issues that are beyond the scope of this evaluation (Glantl, 1987). Meteorological drought is usually defined in terms of a period of a year or more with a large negative rainfall anomaly. Agricultural drought is such a negative anomaly that impacts significantly on agricultural production.

15

Impacts National & Intemational Responses Southern African Drought

Table 2.1 Index of Cereals Production in Southern Africa, 1991-94

Marketing Year

Country 1991/92 1992/93 1993/94 (p)

Severely Drought Affected: 1986/87 - 1990/91 = 100

Botswana 83 25 128 Lesotho 59 41 91 Malawi 112 47 60 Mozambique 94 43 137 Namibia 125 36 125 South Africa 88 39 104 Swaziland 110 39 53 Zambia 80 39 69 Zimbabwe 89 22 125

Total 90 38 112

Other SADC: Angola 116 116 122 Tanzania 93 89 93

SADC Total (a) 93 57 95

Source: SADC Food Security Bulletins, FAO and IWC Notes: a. Excluding South Africa

b. Provisional, March 1994.

2.2.4 Estimates of import requirements for the severely affected SA DC countries rose from the 1.6 million tonnes level during 1991/92 to over 5m tonnes for 1992/93 by April 1992. Taking into account the response measures considered below, actual food imports, including both concessional and commercial flows, of the nine seriously affected countries, including RSA, increased during 1992/93 by 7.2m tonnes or 235%. RSA's agricultural sector was severely affected and the country had to abandon its traditional role of exporting to the four participating countries within the Southern Africa Customs Union (SACU) and import on a large scale. Cereal import requirements were in excess of 4m tonnes in landlocked Malawi, Zambia and Zimbabwe, whilst the requirements of all the other countries in the region rose, in particular in Mozambique where the drought intensified the impact of the distressful continuing civil war. Inevitably, regional and international attention focused on the cereal import problem which, unless met, threatened widespread shortages and potential famine conditions. None of the countries maintained stocks that could do more than cushion temporarily the impact of such a severe shock and even the regional food security arrangements under discussion amongst the SADC countries did not envisage inter­year storage on this scale.

2.2.5 The direct impacts of the drought on food production were aggravated by wider imRacts on agri cultura 1 and non-agri cultura 1 production and on export revenues.' The latter were concentrated in Zambia and, more particularly, in

For ao anal'ysis of the wider economic and financial consequences of drought with case slHdy material from southern Africa in 1991·92, sec Senson and Clay, 1994.

16

Southem African Drought Impacts National & International Responses

Zimbabwe, where power generation and industrial production were affected directly by water shortages and indirectly by the recessionary impacts of the slump in agricultural production. Consequently, all the countries were confronted with both internal and external financing gaps, to pay for massive increases in cereal imports and measures to facilitate agricultural recovery as well as the responses to the water- and health-related drought effects. Moreover, such responses had to be made in the context of the wider negative economic consequences resulting from relative or absolute reductions in economic activity. Current trade accounts in several countries were negatively affected by the decline in agricultural and non-agricultural exports affected by the drought. Governments were also confronted with reductions in their revenue accounts at the same time as facing additional drought-related expenditures.

2.2.6 Overall, the impacts of the drought on the peoples and economies of Southern Africa were complex and differentiated as is shown in the three annexed case studies. Impacts are only partly explained by the severity of the drought in meteorological or agricultural terms. The human and economic consequences depended on the structure of the affected economies and the specifics of their immediate situation. In terms of direct economic impacts, probably Zimbabwe and Zambia were the most affected, with possible disruption in the processes of economic restructuring in Zimbabwe and political liberalisation in Zambia. Malawi was also severely affected by an interaction of the direct impact on the economy with severe internal economic and political difficulties, as well 8.S the presence of over 1 million Mozambican refugees. Mozambique was arguably the country most likely to experience the early development of famine conditions (see Annex 1). In Botswana, Lesotho and Namibia the impacts appear to have been primarily on directly affected rural populations, the macroeconomic developments being disturbed only to a limited extent. In aggregate terms, South Africa's agricultural sector was the most severely affected.'

2.2.7 For the governments of the region and the international community confronted with a complex disaster, the principal features were the huge increase in import requirements, particularly for landlocked Malawi, Zambia and Zimbabwe; Mozambique's increased relief requirements; and South Africa's unprecedented commercial import requirement of over 5m tonnes. The logistics involved in such a major importation programme, which required the movement of cereals in the opposite direction to the normal flows out of Zimbabwe and South Africa, became the main focus of the international response.

2.2.8 Furthermore, there were complex and differentiated relief problems to be addressed in countries with very varied capacities to mount large-scale operations. Some with smaller populations, such as Botswana and Lesotho, already had in place food distribution systems organised around drought relief. In Mozambique existing conflict-related relief systems needed to be expanded to cope with the impacts of the drought. Zambia and Zimbabwe were confronted by temporary relief problems as well as by their large economic ones. Relief problems were most severe in the more arid areas that had already experienced localised drought in the preceding years. Drought posed problems of organising recovery, especially for the worst hit small-scale farmers and water supply and health problems were also potentially severe. Finally, economic impacts were highly differentiated. An important issue, therefore, in assessing the ODA's response in the context of the overall international one is the extent to which ODA was sensitive both to the drought's differentiated consequences and to the need to ensure that its assistance complemented other actions.

South Africa's cereal production, which normally counts for about two·thirds of the total in the affected countries, was reduced from around 12m in the late 1980s to 4.7m tonnes, and related imports rose from under half a million to over 5rn tonnes during 1992193.

17

Impacts National & Intemational Responses Southern African Ormlgl1t

2.3 THE RESPONSE BY GOVERNMENTS AND THE INTERNATIONAL COMMUNITY

2.3.1 The full severity of the crisis did not become apparent at any single moment. Instead, a crisis of ever more serious proportions gradually emerged as the failure of the rains became more complete and crop production prospects deteriorated. It is this aspect of drought as a slow-onset disaster that poses particularly challenging problems to national and international institutions.

2.3.2 As a result of improved information systems, institutions are not now suddenly confronted with a drought. The problem is pre-eminently one that calls for a sequential and flexible response. Where food imports have to be organised either commercially or as food aid, there are likely to be severe risks of a late response if deci si ons are deferred unti 1 the full scale of requi rements bp comes apparent. Similarly, as local food production recovers, water supply improves and related health risks subside, there is a need to modify responses. There are then risks of over-supply and negative impacts on local agricultural recovery. There is then scope, however, for redirecting the response from relief measures to actions that will facilitate a more rapid and sustained recovery. These aspects of a severe drought shock provide criteria for assessing the response of the international community and, within that context, the actions of ODA.

2.3.3 As early as August 1991 the FAO's GIEWS and the SADC Food Security Technical and Administrative Unit (FSU) had cautioned that the combination of a relatively poor harvest in 1991/92 and Zimbabwe's maize export commitments would reduce regional cereal stocks below safe levels and probably necessitate additional imports in the first half of 1992, but no action was taken. During December 1991, the US Government began to indicate its serious concerns internationally, for example at the Zimbabwe Consultative Group meeting in Paris. By the end of January 1992 the GIEWS and SADC's FSU had accumulated convincing evidence that the region was suffering an unprecedented drought, and in early February the geographical spread of the impact was starting to become clear. On 12 February President Chiluba of Zambia declared the southern, western and eastern provi nces of hi s country to be di saster-affected and two weeks 1 ater widened this to the national level. The GIEWS Alert of 21 February 1992 drew attention to a food crisis of regional proportions. This report was tabled as the main background document for a meeting of all of the major food aid donors in Brussels four days later which recognised the gravity of the crisis and set in motion the process of full assessment and response.'

2.3.4 On 25 February FAO presented estimates of cereals production in the region of approximately 25% below normal, with import requirements of 3m tonnes for South Africa, 1m tonnes for Zimbabwe and 1m tonnes for Mozambique. FAO and WFP jointly undertook to make individual country food supply assessments, which were comp 1 eted between 23 Ma rch and 10 Apri 1 1992. The need for a regi ona 11 y co­ordinated logistics programme was recognised, with WFP again taking the lead, and the US undertaking additional logistical activities for South Africa.

2.3.5 The GIEWS Alert of 15 April based on the FAO/WFP assessments estimated aggregate import requirements of the 10 SA DC countries as 6.1m tonnes, some 1.6m tonnes of which was to be emergency aid for direct distribution, with the remainder to be provided as programme aid or commercial imports (see Table 2.2). The estimate of programme food aid requirements attempted to take into account the exceptional commercial import programmes already organised by Zambia and Zimbabwe and the normal commercial imports of Botswana, Lesotho, Namibia and Swaziland. It reflected in part those additional imports that countries in the

The meeting on 25 February was part of an irreglllar series of informal gatherings of food aid donors. It was fortuitous in its timing, and not organised specifically to consider Southern Africa.

18

Southern African Drought Impacts National & International Responses

region were looking to donors to fund directly as food aid.' The total food aid requirements of over 4m tonnes were very large in relation to global food aid supplies which had averaged 10-12m tonnes annually since the mid-1980s. They therefore implied some combination of additional allocations during this and the next financial year with some possible diversion from other recipients.

2.3.6 The UN Department of Humanitarian Affairs (DHA) and SA DC agreed to make a joint Drought Emergency in Southern Africa (DESA) appeal, eventually scheduled for the end of May 1992, prior to an international pledging conference in early June. But by that time the governments of affected countries had begun to organise their own responses and donors had already formulated and begun to implement plans of action. ODA itself had made its initial commitments as early as 3 March. The United States had begun to put together large packages of food aid and agricultural export credits in February-March 1992. In April, the World Bank agreed emergency recovery credits for US$150m for Zimbabwe and US$100m for Zambia including financing of food imports and the European Community agreed a package of drought assistance on 4 May.

2.3.7 The assessment phase had inevitably been an extended process and the FAO/WFP assessments and the DESA appeal had added over three months, to confirm the dimensions of what was clearly the largest-scale food crisis in Africa. Both the governments of the affected countries and donors were acting responsibly in beginning to organise their responses, if on a partial basis, prior to the DESA appeal. Indeed, in the attempt to achieve better overall co-ordination, the process of formal assessment and preparation of a regional appeal may have contributed to delays in the response by some donors.

2.3.8 A new and constructive feature of the international assessment process was the attention paid to non-food aspects of the drought and to non-food assistance requirements. The DESA appeal requested some US$208m for the health (and nutrition), water, and agricultural sectors as well as for targeted food aid and logistics, institutional support and multisectoral support for displaced people. By the end of 1992 US$138m, or approximately two-thirds, had been contributed. On the basis of this figure the international community had apparently responded inadequately to the non-food aid requirements of affected countries (DHA/SADC, 1993). The greatest apparent shortfalls occurred in agricultural rehabilitation (only 7.3% fulfilment), displaced people (22%) and water (32%). This apparent gap, however, between needs and response cannot be interpreted in any straightforward manner, and requires closer examination.

2.3.9 First, the non-food aid elements of the appeal represented only the proposals of UN specialised agencies and some of the components were not strictly concerned with either immediate emergency requirements or rehabilitation, and so imply an over-estimation of emergency needs (Borsotti, 1993). Second, some of the country proposals failed to take account of bilateral donor and government actions in the same sectors and so involved an element of duplication. Third, the non-food assessment was seriously incomplete because components were included only if they were part of proposals being made by specialised agencies. The DESA did not monitor systematically donor (including NGO) responses, apart from those directly concerned with its own components. Even an incomplete review indicates that the non-food response was considerably greater than was implied in DHA/SADC reporting. For example, some of the ODA bilateral assistance, including that channelled through NGOs, was not fully reported. Nor was the large World Bank response fully reported.

For example, Ihe programme aid requirements for Zimbabwe were quite modest, some 180,000t. Tanzania, hardly affected by the drought. had programme aid requirements of over 300,0001. Mozambique, with a Pllblic distribution system already stretched 10 capacity, had projected reqllirements of over 700,0001.

19

1mpacts National & International Responses SOllthern African Drollght

Table 2.2 Southern Africa: Cereal Import and Food Aid Requirements for 1992/93 ('000 tonnes)

Import Requirements

Country Food Aid Requirement Total

Total Of Which: Food Aid as Emergency a % of

Total

1. SADCC

Angola 285 125 116.0 44 Botswana 240 15 3.3 6 Lesotho 297 75 11.2 25 Malawi 876 740 400.0 (b) 84 Mozambique 1,218 1,140 414.0 (b) 94 Namibia 125 ( c) 60 16.5 48 Swaziland 129 60 41.0 47 Tanzania 530 327 (15.0) 82 Zambia 970 820 94.0 85 Zimbabwe 1,413 650 470.8 46

Total SADCC 6,083 4,012 1,582.0 66

2. South 5,500 ----- ------- --Africa

3. Drought 10,768 3,560 1,451 33 Affected

Countries (a)

Source: FAO/WFP Mission, April 1992 Notes: a. SA DC less Angola and Tanzania plus South Africa

b. Includes requirements for Mozambiquan refugees in Malawi c. Net imports

2.3.10 The implications of the apparent under-response are unclear. For example, the largest sectoral shortfall was in agricultural rehabilitation, but regionally there was a rapid recovery in 1993/94 (Table 2.1). However, little is known about the extent to which the most severely affected groups in particular countries were able to participate fully in that recovery. If some severely affected groups, such as agriculturists or pastoralists in low-potential environments, were further marginalised, then problems associated with poverty and vulnerability to future crises may have increased.

20

Southern African Drought Impacts National & International Responses

2.3.11 Turning now to the responses of affected countries, in late 1991 the South African authorities began to plan for a large-scole increase in food imports, and by mid-January they had frozen further maize exports outside the customs union (SACU). This had serious implications for Zambia, Zimbabwe and Mozambique. Zambia, already facing severe economic difficulties during a phase of political liberalisation, responded most quickly of all the SADC countries. With informal evidence of a major disaster accumulating, the Ministry of Agri cul ture began to take emergency measures and the Presi dent decl ared an emergency on 12 February 1992. The responses of three severely affected countries are covered in detail in the case studies. In summary, Zimbabwe, after being initially slow to recognise the scale of the crisis, began to arrange commercial imports and informally seek assistance for food imports in February 1992. After the declaration of a national emergency on 6 March the government responded decisively in organising food imports and a national drought response programme, both on an unprecedented scale. The increasing severity of Mozambique's food situation was recognised early, but the effects of drought were interacting with the continuing confl ict-related emergency. The government therefore sought assistance that included all aspects of the complex emergency. Because of the government's limited implementation capacity, international NGOs and the UN agencies played major roles in assessing the needs of affected people and requirements for implementable relief activities. Lesotho's government responded comparatively slowly; a full assessment was only completed in July 1992 after donor pressure and intensive advocacy by NGOs.

2.3.12 The scale of the drought necessitated an unprecedented level of imports in the region and a matching regional logistics response. Over 10m tonnes of maize needed to be procured, shipped, landed and transported during the 1992/3 marketing year and this placed an enormous strain on the existing transport and storage infrastructure.

2.3.13 The WFP assessment missions found that there was sufficient port and offtake capac i ty ava i 1 ab 1 e to cope with the es t i mated 10. 8m tonnes of food imports needed, inclusive of South Africa (Table 2.2). WFP took responsibility for monitoring and co-ordinating the shipping schedules and deliveries and the subsequent transport and distribution of all multilateral and bilateral food commodity consignments under its own regional emergency operation and in general through the Logistics Advisory Centre (LAC) in Harare (set up in June 1992 by SADC and WFP). Donors, including ODA, pledged some US$30.2m through WFP to fund this logistical co-ordination.

2.3.14 On 18 June under the Umbrella Drought Emergency Operation for Southern Africa WFP itself initially committed over 700,000 tonnes of relief food from the International Emergency Food Reserve. Following additional donor contributions, this was increased to 845,000 tonnes. WFP handled on behalf of bilateral donors an additional 35,000 tonnes of food aid.

2.3.15 The achievements of the logistical response can only be appreciated if both commercial and food aid imports are taken into account. The combined import requirements of South Africa and all the SADC countries were estimated by FAO/WFP in April 1992 as over 11m tonnes, including 4m tonnes of food aid for the latter in 1992/93 (Table 2.2). In the event, some 10.5m tonnes of food arrived in the countries of final destination during 1992/93, including 5.6m tonnes for SADC countries and 4.9m tonnes for South Africa - an increase of 7.2m tonnes over imports in 1991/92. Especially impressive were the large increases in imports to Zambia, Zimbabwe and Malawi. The volumes of food moved through the ports of SADC countries were also impressive, especially in the context of the continued confl i ct up to 1 ate 1992 in Mozambi que. The enormous increase in food import requirements, especially of the larger landlocked countries, has been confirmed by most evaluations as the crucial aspect of the crisis.

21

Impacts, National & International Responses Southern Alrican Drought

2.3.16 The success of the combined efforts of the governments of the region working with the international community in establishing requirements, making commitments of food and organising the logistics is thus the central aspect of the overall response. The initial assessment of the problem was, at least by comparison with some earlier crises, successful in terms of the speed with which food problems were identified and crucial elements of the response to these problems organised. A second aspect of the response is the adequacy in terms of providing resources in both overall terms and in particular components. The DESA process was both central to and as discussed above, a not wholly satisfactory aspect of this process. Thirdly, the response can be considered in terms of the organisational arrangements for ensuring the provision and arrival of food. Broadly, the regional logistical operation was a success in terms of unprecedented co-ope:--ation and co-ordination within the region involving the drought-affected SADC countries, South Africa and the international community.

2.3.17 In the context of the contribution of ODA to the overall success of the international response several factors coincided to make it likely that the response to the major drought in Southern Africa would be successful. First, the region has effective early warning arrangements for a food system which is fairly well integrated in market terms. Second, the drought had a major impact upon RSA and consequently, through its early decision to begin to import on a commercial basis, upon the international grain trade. Third, as the crisis was likely to have implications for the cereals market, this in turn involved the USA, which played an important role from an early stage. Fourth, an important reason for the relatively rapid and generous international reaction was unquestionably political. The possibility of disruption in the processes of political reform in RSA, as well as those in Zambia and elsewhere in the region, such as Malawi, and of derailment of the structural adjustment programme in Zimbabwe, could not be ignored.

22

3.1 Monitoring: ODA seeks to monitor the situation in the Southern African region on a regular basis via the country desks within CSAD and officials in post in the various countries. Historically, this is a region of considerable British interest, and consequently most Posts have a substantial developmental role in proporti on to the si ze of thei r programmes. ECFAD has its own network of contacts and information and maintained a watching brief on the southern African situation, particularly during the period December 1991-February 1992.

3.2 CSAD and ECFAD jointly concluded, by the end of February 1992, on the basis of information coming in from officials in post, other donors and international early warning systems, that a crisis of severe proportions was developing in the region. In particular, it involved those countries where the UK had substantial developmental and commercial interests, especially Zambia and Zimbabwe. At this stage, EMAD was also involved. The initial internal assessment reflected contact with other donors, and information from the UN system, some of the governments of affected countries and the Posts, and determined the main elements of ODA's strategy.

3.3 Objectives: In determining its initial response announced on 4 March 1992, ODA distinguished two separate but complementary objectives for providing assistance to drought-affected southern Africa. First, it would contribute directly to alleviating the suffering of affected people by an appropriate combination of food aid and emergency aid. Second, affected countries should be assisted in coping with severe balance-of-payments problems, ie economic effects of the drought. It was also recognised that ODA would be contributing to a wider response, and that its bilateral response should be made with a view to galvanising the wider international community, in particular the European Community, into making early and adequate responses.

3.4 Strategy: ODA's approach was to make an initial response as rapidly as possible to assist those countries that were, on the basis of preliminary assessments, the most severely affected. It woul d then avai 1 itself of the results of fuller international assessments before re-examining the appropriate scale and distribution of its overall response. Its initial package of assistance, as reflected in the announcement on 4 March 1992,9 had four elements.

3.5 First, two of the affected countries, Zambia and Zimbabwe, were recognised as facing a serious resource gap in financing large exceptional food and other agriculture-related imports. The drought crisis was seen therefore from the outset as having an economic as well as a humanitarian dimension. It was decided that these problems should have a priority claim on CSAD bilateral development funds remaining for 1991/92 and in 1992/93. Additional financial aid was committed of £2.Sm to Zambia in 1991/92 and of £2.Sm in 1992/93 to Zimbabwe.

3.6 Second, UK bilateral food aid organised by ECFAD was to be targeted on the countries where emergency food aid was considered to be the most appropriate response. The regular programme for 1992/93 would provide 10,000 tonnes of maize for Mozambique and 5,000 tonnes for Mozambican refugees in Malawi (val ued at £3.9m).

3.7 Third, ODA would provide technical assistance where necessary to the regional logistics effort. The role for emergency assistance administered by EMAD was not at that stage clear, but it was expected that some response through NGOs was likely.

ODA Nt''''I IIdm.l<', 23/')2, 4 "br~h l'i'l3

23

OOA's Response Southern African Orought

3.8 Fourth, having decided to make a substantial response, the strategy would be proactive in encouraging and facilitating European and wider donor responses to the drought. The UK's own early bilateral action was intended as a demonstration of its assessment of the seriousness of the situation and its commitment.

3.9 Phased response: ODA decided to await fuller information from the FAO/WFP assessment and from officials in post before taking further action, although it responded on 11 March to a WFP appeal for urgent assistance to Malawi with £lm. for targeted feeding. Its own internal assessment on 20 March, which had involved consultation with international agencies in Rome, confirmed the gravity of the food crisis. The size of the projected food import gap had increased by approximately SO% within a month. The budgetary implication regionally was some US$lbn.

3.10 A meeting with the principal UK NGOs was held on 30 March, in the expectat i on of subs tant i a 1 requests from Brit ish-based NGOs to support thei r relief efforts. There were also a number of international consultations, including those organised by the World Bank. With evidence of an ever more serious crisis, ODA determined on a further package of measures, involving intensive consultation with officials in post, on logistical constraints, food distribution arrangements, donor co-ordination and NGO activities.

3.11 A second tranche of assistance was announced on 29 April comprising: a further £7.Sm for Zambia and a further £Sm for Zimbabwe in terms of financial support; £Sm for food aid mainly for Mozambique; £2.Sm for other drought-affected countries in the region, including Botswana, Lesotho, Swaziland and Namibia; support for the activities of British NGOs; and technical assistance as needed. This second response consolidated the action by CSAD and those in post to identify needs and opportunities for additional bilateral funding. It provided the additional dimensions of a complementary response involving British NGOs. ODA also offered some assistance to regional co-ordination, in particular for the Logistics Advisory Centre and the operations of the DHA. The main elements of ODA's response were thus in pl ace pri or to the European Development Counci l' s approval on 4 May of a special additional food aid programme for Africa. By this stage the World Bank, the USA and some other major donors had also made their main commitments.

3.12 With additional activities identified, a third tranche of assistance was announced on 30 June, comprising further balance-of-payments support to Zimbabwe (£2.Sm); additional food aid for Malawi and Mozambique; contributions to UNICEF and WFP regional programmes; as well as £lm of additional support for UK NGOs.

3.13 Overall response: The final total reported commitment of UK bilateral aid up to the end of 1992/93 amounted to £S6m. This included: £10m of financial aid each for Zambia and Zimbabwe to help ease the burden of food imports; £Sm for WFP feeding operations in Malawi, plus a further £4.5m for other drought-related assistance; 10,000 tonnes of food aid for Mozambique and 5,000 tonnes for Mozambican refugees in Malawi (value at £3.9m),'O plus £5m for further shipments of food intended mainly for Mozambique; £5m in technical assistance and support for the activities of British NGOs in the region; £8m for drought-related assistance in Mozambique; £2.7m reallocated for other drought-related purposes from country programmes; and £1.5m for UNICEF/WFP operations in Botswana, Lesotho, Swaziland and Namibia". Closer examination indicated that the reported commitment of £56m included some assistance to the complex emergency operations

Of the 10,000 tonnes committed for M02ambiqlle, 3,500 tonnes was actually from FY 1991(92 funds and is therefore not included in the detailed assessment.

OOA News Refease, May 1993,

24

Southern African Drought aDA's Resronse

in Mozambique that could not be attributed specifically to the drought response. Therefore for the purposes of review, this evaluation concentrates on a total actual expenditure of £49.2m that can be regarded unambiguously as the drought response. The effectiveness and efficiency of these elements of the drought response are considered in more detail in Chapters 4 and S below.

3.14 The phased approach adopted by ODA simplified a process involving three of its departments and officials in post, and was in turn set in the context of its complex relationships with international organisations which were seeking to co­ordinate the overall donor response, the direct appeals of governments of affected countries and the increasingly active NGO community. Such an approach has been advocated si nce the Afri can emergency of 1984/8S as preferabl e to holding back until all the information is available.

3.1S Assessing ODA's response: To assess the implications of these commitments, three complementary statistical analyses were undertaken: on UK aid expenditure in the region as a measure of the aggregate resource flow implications; on the composition of the emergency assistance committed in terms of channels of assistance, i.e. direct bilateral assistance to governments and indirect assistance through NGOs and international organisations; and on the composition of emergency assistance to the three case-study countries.

3.16 Regional Resource Flows Total bilateral assistance to the drought-affected countries amounted to £152.4m in FY 1992/93, a £30.5m or 25% increase on 1991/92, but only slightly higher than in 1990/91. Total assistance subsequently declined by £21.6m or 14% in 1993/94. Table 3.1 shows the breakdown into the various categories of aid in the period 1988-94. Overall, patterns of actual drought­rel ated expendi ture were broadly consi stent with commi tments. It is more difficult to distinguish the element of additionality in total assistance resulting from the drought. There is no clear trend in total expenditure up to 1993/94, suggesting that total aid to the seriously affected countries would probably have been between £120m and £130m in 1992/93, had there been no drought.

3.17 Project aid expenditure has been decl ining apart from the effect of exceptional ATP disbursements in 1990/91 and 1991/92. There may have been some small disbursement difficulties in shifting resources to meet drought requirements in 1992/93.

3.18 Programme aid to Zambia increased from £12.5m to £27.5m between 1991/92 and 1992/93, and to Zimbabwe by £lSm to £22.Sm. In the absence of a drought, however, ODA was already committed to providing programme assistance to Zambia in the context of political and economic liberalisation and to Zimbabwe in the context of the Economic Structural Adjustment Programme.

3.19 Disaster relief to the region rose by £10.6m in 1992/93 and continued at a higher level than in 1993/94 before the drought. The main recipients were Mozambique (£7 .5m), treble that level of 1991/92, Malawi (£4.2m) and Zimbabwe (2.2m) out of a total of £16.2m for the drought-affected region. Botswana, Namibia, Lesotho, Swaziland and NGOs in the RSA also received ODA disaster relief for the first time in at least four years. In addition, ODA contributed £503,000 to Malawi and £100,000 to Swaziland through UN agencies.

3.20 Total bilateral food aid expenditure for the region increased from £3.6m in 1991/92 to £8.3m in 1992/93. Mozambique accounted for £7.6m of the total, Malawi. In addition, Malawi receiving the balance also received £4m of food aid channelled multilaterally through WFP, this sum representing 30% of the ODA contribution to WFP in 1992/93.

25

DOA's Response Southern African Drought

3.21 Taking into account the difficulties of counter-factual analysis, the element of additionality appears to have been between £20m and £30m, out of the total actual £49.2m of drought-related expenditure in FY 1992/93.

Table 3.1 ODA Bilateral Aid Expenditure to Drought-affected Countries in Southern Africa, 1988-94, £m(a)

ril1~nci~1

Ycnrs

J<illllilN 1<)!l<)f')O

1'1')0/<)1

1991191 1<)I)2N3

1<)93N4(c')

Source: Note:

Pwjc<:!

Ai<.l Ill)

l'f(lgm<l1mc Ai<.l

IR,1l

IR.1l S.'l.7

2~5

(,IHI

5'1.1,

Non-

EnICf[!,:ncy

F""d Aid

5.0

r""tI Aid (c')

2.' 1.<)

" 3.6

1\.3

4.7

l)i,.l'!cr Hdid

En1(:rg~I1~Y Debt Aid (d) Rdid(c)

S.l IUI.'l~

('5 (I.!I(l~

7.' 0.111)3

(,,1

1(»<

11.1 0 .. ,114

British Aid Statistics, London: ODA, annual

0111,:r IT. T,,!~r

'4 11111,1

1.3 '1~,1

1.3 1,I'll'

I.J 121.9

2.2 1~2A

2.1 131l,~

a. Includes Botswana, Lesotho, Malawi, Mozambique, Namibia, South Africa, Swaziland, Zambia and Zimbabwe

b. Includes financial aid, technical co-operation and Aid and Trade Provision.

c. Includes 'directed' contributions to WFP. d. Includes Disaster Relief, Refugee Relief and Disaster

Preparedness. e. Provisional.

3.22 Channels of ODA response involved 43% to governments, 36% indirectly through NGOs and 17% via international agencies (Table 3.2). The direct aid to governments was largely accounted for by financial support to Zambia and Zimbabwe. Otherwise, a limited amount of emergency and technical co-operation funds were di rected to governments and moni tored by Posts. Food ai d was channelled entirely through NGOs in Mozambique, and through the WFP in Malawi. In addition, a limited amount of further support was provided to WFP regionally for local purchases in Botswana, Swaziland, Lesotho and Namibia.

3.23 The emergency aid provided indirectly through UK-based NGOs and the assistance channelled through international agencies were spread more widely across the region, but were still heavily concentrated in Mozambique, Malawi and Zimbabwe. The choice of channels appears to have taken into account both ODA's assessment of impacts and the response of other donors.

3.24 The sectoral composition also reflects in part priorities determined internally by ODA in its initial response and its responses to proposals from governments, NGOs and international agencies (see Table 3.3). The largest single element of assistance, balance-of-payments support, reflected ODA's assessment of the severe economic impacts in Zambia and Zimbabwe. In contrast, for Malawi and Mozambique the preferred response was food aid for direct distribution. These decisions reflect at least implicit judgments about the capacities of national institutions in the drought-affected countries. The other elements of the response to a large extent reflect the balance of proposals put to ODA by NGOs and international organisations, directly or channelled through the Posts, for using emergency and bilateral technical co-operation funding. Water had relatively higher priority in Lesotho, RSA, and Zimbabwe, agricultural rehabilitation in Malawi and Mozambique. Support for relief food distribution (including logistics) was spread broadly across the region. Health was in

26

Southern African Drought ODA's Response

Table 3.2: ODA Bilateral Response to the 1991/92 Southern Africa Drought by Channel of Assistance (£'000)

i. Assi~lal1,:e in £'000

B('I!.\v~na

L"~\'lh" Malawi

I-lwamhi'l"c Namibia S,'ulh Ar[i<'~

S",,,,il'lI1d

Ziml>;,hwc

Rcgillnal

To!al

Ch'lrllld a~ '/,. "I' 101.11

all"calillll

Source: ODA

HI.OIHI

111,(11111

20.01HI

.1(1.7

vi:1

[\'sl

JH5

145

2('7

51111

1..,1111

2.(,

Indired

NCO

22.1 4'1H

.1.(,.1'1 (a)

3,X42

·LW 7.17 DI H2.'

1..152

I1.HI(,

2~.!I

Inkrn:.li"nal

"t!C'n()'

~2()

1.17.' (I) son 25·1

" .'J.' (a) 2')5 ,'{(d <I»

.1.7'/.1

5.7

j\'GO

(,,SOH

(,,5I1H

\.\.2

Inlanalinllal

AFncy

5.0110

HI,IHIO

7."11

5.750

! 1.7

T"lal

223 1.212 'I.leH

I(I,H.~(I

71J HH2 17·1

11,·12S

12.2·17

1/,1.1

~'}.If>H

!!lIi.(IIJ

Notes: a. The precise channelling of £O.721m to Malawi and £O.319m to Zambia could not be determined on the basis of available information.

b. Includes contribution to DHA Geneva.

general a low priority, with proposals coming largely from WHO and UNICEF for indirect support to national programmes.

3.25 The important issue for assessment therefore is the relative effectiveness of those allocations reflecting ODA's choice of priority and those where it was reacting to NGO and international agency requests. This balance of priorities al so has to take into account ODA' s awareness of other donors' acti ons. The European Community and the US were responding primarily through the provision of food aid. Aid consortia discussions and other information channelled back from Washington indicated that the World Bank, the African Development Bank and some other bilateral donors were providing additional finance for drought-related imports as well as re-allocating support initially provided to support economic reform.

3.26 The specifics of ODA's response are, however, best understood at a country level, and the three case-study countries described more fully in the Annexes illustrate the range of circumstances.

3.27 In Zimbabwe ODA, in consultation with the GoZ, was highly proactive in determining the overall response. There is an interesting opportunity in this case, therefore, to contrast the performance of those major elements of support that reflected ODA decisions which also involved the Post, (namely the provision of financial aid (Section 4.2) and bilateral project aid for the water sector (Section 4.4)), with the performance of those grants essentially reflecting decisions by British NGOs and their local partners.

27

ODA's Response Southern African Drollght

Table 3.3 ODA Bilateral Response to 1991/92 Southern African Drought by Recipient Country and Sector (£'000).

(,<>""If)"

i. EXl'c'nti;lurc

in £ Ih""~nTld

l3"hw~na

us,,!ho Malawi

M<>/al1lhi 'luc Nmnihia S"uth Ai"ri<:a Swni!aml

Zilllh;li,wc

R<:gi<>nai

To!;,)

ii. S<'<'I<\r a.~

~I "I' \(\tal

Source: Notes:

n.II"",<: ,,'

f"lYIll~!II'

!O.lIOIl

IIl.lIOO

20.iJ(lO

40.7

ODA

W:llcr

22<)

45

10·)

3.0(,1

Ikillh

20<)

50.\ 77

I.Ol'l.!

2.2

'\1= 1'<1,,<1 r""d Hd,.oI, Aid 1)','ril>·

"uun

~'J3

3l'11 33il (,.\

3ill)

" 1 J\} 7g2

75(]

12.2.'l'I 2.755

'1.1 5.f>

I"r;""!,,>rl ~\"llj· \""1.11

" .'<"c·l"r

1.<'!!;Slj,,-\ (al

22.1 223 2~ 1.212

1,0\3 'I.l'I2l'1

1.i12~ 31111 1{t.RSO (h) 21'1 37<) 7!.l I)R 13.~ RR2

25 20 )7.j

33." (\17 11.42.' 21K 12.247

!\(,.! 1/1\4

I.<)x~ )557 .\<).16x

~.I! 7.J \11I1.1!

a. Some expenditure that could not be specifically attributed to a particular sector: £752,000 in Malawi, £319,000 in Zambia and £435,000 regionally is assumed to be multi-sectoral.

b. Including all actions defined by ODA as drought commitments, drought rel ated acti ons under regul ar emergency commi tments and Rome pledges, and Unicef £0.5m water project.

3.28 Mozambique is sui generis because of a continuing emergency situation, with a number of UK-based NGOs already involved in relief operations. Having made major decisions to target food aid and additional drought-related assistance to Mozambique, the actual grants in terms of sectvrs, as well as reg,unal targeting, reflect these eXisting realities (Section 4.3).

3.29 Lesotho is one of the smaller drought-affected countries where ODA responded to government requests for assistance as well as providing support indirectly through British-based and international organisations. There is an opportunity, therefore, to contrast the performance of a bilaterally supported governmental programme in the water sector (Section 4.4) with indirect support for national programmes through NGO and international agencies.

3.30 Finally, ODA's response has to be set in the context of the wider international response. In view of the overall food aid response, ODA correctly concluded, having committed its own bilateral food aid, that financial resources could be used in ways which complemented this aid. The issue for assessment is whether or not the combination of choices by ODA, and its responses to proposals from governments and the NGO community, were appropriate and timely.

28

4.1 CRITERIA AND METHODS USED

4.1.1 In examining the various ODA-supported components of the drought response, the following distinct areas of activity can be identified:

• balance-of-payments support to governments of affected countries in financing import programmes associated with the drought;

• food aid, relief distribution and logistics. In practice, it proved difficult to separate these into the various components of food rel ief operations, including the acquisition of food and equipment, direct delivery costs and management, covered by grants;

• emergency rural wa ter programmes;

• agricultural rehabilitation measures, particularly seed distribution;

• health measures.

4.1.2 The evaluation's assessment of the appropriateness and effectiveness of ODA's response considers the following aspects of actions in these sectors:

• the selection of channels, including direct support to governments and operational NGOs, indirect support through UK-based NGOs to locally operational NGOs, and support to NGOs mediated through government distribution channels;

• the design and targeting decisions of agencies involved in the drought response and overall of ODA in terms of the balance of countries and agency programmes supported;

• the forms of ODA support to whole programmes or to components of programmes by means of financial aid, imported food aid and grants to cover purchases of food, seeds, drugs, equipment etc., as well as organisational and management activities;

• evidence of the relative efficiency of ODA and its partner organisations in terms of response times and the appropriateness of the components supported;

• monitoring and evidence of impact on final beneficiaries.

4.1.3 The methods of study adopted by the evaluators were as given in Chapter 1, with individual variations as spelled out in the three case studies. In the case of emergency water projects which usually result in physical structures, there were site inspections and analyses of primary, technical and financial information. For the other components, the evaluation is based on information resulting from the monitoring and assessments by grantee agencies, as well as from government and other organisations involved in the drought response. The only practical alternative was to organise social surveys of intended and actual beneficiaries to obtain recall information. However, this study was not organised to allow for what would have been a relatively costly collection of information. In any case, with so many ODA-financed actions, findings fr0m a few surveys would have been illustrative rather than representative.

4.1.4 ODA does not make formal assessment of impact a requirement of grants, and does not usually provide funding to assist this activity when making emergency grants. Nor did the majority of the organisations concerned, whose emergency

29

Impacts in Terms of ODA's Response So\!lliern African Drought

actions were in any case supported in most cases only partially by ODA, make performance monitoring and impact assessment priorities. As a result, the information available on impacts is inevitably patchy. A few organisations involved in the drought response did give priority to monitoring and assessing impacts and their evidence is used to illustrate issues in impact assessment. Otherwise, information is at best fragmentary and qualitative.

4.1.5 The broad picture at a regional level is relatively clear. As indicated in Chapter 2, the combined response of governments supported by the international community prevented what could have been an economic and human catastrophe on a considerable scale. The precise impacts, however, of most ODA-supported actions at a community level and on the status of specific groups, such as women, children and potentially disadvantaged minority ethnic groups, is uncertain.

4.2 BALANCE-Of-PAYMENTS SUPPORT

4.2.1 The provision of balance-of-payments support in the form of additional £10m grants to the governments of Zambia and Zimbabwe to finance drought-related imports, was the largest single form of support in ODA's response, 40.7% of the £49.2m bilateral action. These actions were found to have been important positive actions by ODA and exemplary in their implementation. To those seeing the drought primarily as inducing a food crisis threatening potentially grave human consequences, this may seem surprising. If so, this may be because the potentially large economic cost of drought shocks to affected countries has received little attention in the past (Benson and Clay, 1994) (see Annex 2 on Zimbabwe).

4.2.2 Financial assistance, as well as direct material assistance where this substitutes for expenditure that a government woul d otherwi se have incurred, di rectly reduces the domesti c fi nanci ng requi rement of a drought emergency programme. The rapid disbursement of financial assistance is, therefore, especially important in counteracting the recessionary effects of a drought shock.12 The use by ODA in Zambia and Zimbabwe of the arrangements already in place for attributing additional financial assistance to existing import programmes, was important in this regard. This is confirmed by the disbursement problems encountered by IDA credits provided under World Bank emergency projects in Zimbabwe linked to additional drought-related imports and local currency expenditure, in accordance with IDA tendering procedures. 13

4.2.3 The actions of ODA, along with those of the World Bank and other bilateral donors, provided crucial support to these governments in their prompt actions in organising imports to feed their own peoples. The actions had two aspects. Aid initially programmed to support reform was allowed to be used for drought-related imports. Subsequently, ODA and then the World Bank provided additional BoP support. Collectively these actions contributed to mitigating the impact of drought on economic activity by minimising the crowding-out effects on other essential imports of raw materials, inputs and spare parts.

4.2.4 The major criticism that can be made retrospectively is that the totality of additional financial support from all sources to Zimbabwe and also to Zambia was inadequate. Whilst approximately covering the cost of imports directly related to the drought response, this aid covered only part of the balance of payments gap caused by the drought shock as this had also sharply reduced export

"

"

Financial assistance reduces government's financing requirement to be met by borrowing, additional taxation measures or de licit linancing, which may aggravate the negative effects of the drought by crowding out private sector borrowing and private consumption expenditure.

For example, lOA credits provided as part of an Emergency Drought Recovery and Mitigation Project {EDRMP) have to be set against specific internal or external procurements that have to be made according to normal World Bank tendering procedures. Consequentfy, commitments made in April 1992 had only been partially disbursed in Zambia {38%) by the time aOA grants had been wholly disbursed in December 1992.

30

"

Southern African Drought Impacts in Terms of ODA's Response

revenues. The reces s i onary effects of the drought were on 1 y mit i ga ted. In terms of initial objectives and targets, there was a substantial setback to the process of economic reform. Medium-term economic prospects were also negatively affected by drought-related increases in both short- and long-term foreign debt (World Bank,1993).

4.2.5 The question arises: would the provision of balance-of-payments support have been appropriate elsewhere? Although both Malawi and Mozambique required considerable additional imports, these involved a large proportion of targeted relief to drought-affected and war-displaced people; thus making a food aid­related response more appropriate. Nor was large-scale BoP support appropriate for the other countries with relatively small populations, where arrangements were not in place to provide such support (Lesotho, Swaziland); where the drought impact was more localised (Namibia); or where there was no major resource gap (Botswana) .14

4.2.6 Lessons: in what circumstances might provision of financial support to a disaster-affected low-income country be appropriate in the future? Such a response is appropriate where there is:

agreement between donor and recipient government on overall aid and public expendi ture pri ori ti es and the reporti ng requi rements are sati sfi ed by arrangements that are already in place. It is more appropriate where the major response is by the government itself, and there is a related resource gap that would not be filled by existing international compensatory arrangements.

Directly assisting a government's own import programme should have priority where this complements and underpins relief efforts by local and international NGOs. As the Zimbabwe case study illustrates and evidence from Zambia confirms (Seshmani, 1993), NGOs were only able to launch their relief operations in a relatively timely way and sustain these activities because they could draw upon marketing channels that were sustained by the government's own import programme.

A preference for fi nanci al assi stance mi ght al so be conti ngent on the pragmatic judgement that commercial imports are 1 ikely to provide a speedier response than conventional food aid.

ODA needs to be satisfied that the government has an appropriate policy framework, with acceptable monitoring arrangements for reasons of accountability. This is most likely to be the case where the government is already implementing a structural or stabilisation programme in agreement with the international financial institutions.

As the subsequent assessment of other components of ODA IS assi stance implies, the attempt to channel substantially more aid for relief or other drought-related measures in southern Africa in 1992/93 though indirect channel s mi ght have encountered probl ems of di sbursement and cost-effecti ve utilisation.

The WFP evaluation draws similar conclusions, Food aid supplies take tilree,six months to reach a country alter assistance has been approved, Commercial imports by the government can provide a speedier response il the foreign-exchange constraints can be Ilvercome. The contrasting impact of speedy actions taken by Zimbabwe and Zambia on the one hand, and the slow and halting decisions in this regard in Malawi, provide a useful guide for future action. The example of the UK GOvernment in Zimbabwe where financial assistance fllr non· food imports fCleased foreign-exchange reserves in the country for commercial food imports also deserves serious consideration by donors (WFP, 1994:17).

31

Impacts in Terms 01 DDA's Response Southern African Drougllt

4.3 FOOD AND LOGISTICS COMPONENTS

4.3.1 The provision of emergency food relief included:

oD food aid from the bilateral programme for rel ief in Mozambique and for Mozambican refugees in Malawi, and finance for local purchase operations in the region by WFP;

.. grants to NGO rel ief operations, variously including local purchases of food, organisation of relief, storage, handling and transport costs up to delivery to final beneficiaries, equipment for relief operations (largely vehicles), and financing the management of emergency operations including technical support to public relief operations.

4.3.2 These various forms of support for food aid, distribution and logistics operations accounted for some £17m out of £49.2m overall (35%) but, taking into account the food relief components of multi-purpose grants, probably represented at most 38% of ODA's total bilateral response (Table 3.3). The following questions about the overall appropriateness of ODA's response are considered in the case studies:

As ODA was using financial resources to support emergency activity, was there any evidence that particular components of that support - imported food aid, financing of local purchases, transport and handling costs, equipment or the organisation of operations - were relatively more appropriate?

Were there discernible advantages or disadvantages in direct support to opera t i ona 1 agenci es compa red wi th i ndi rect ass is tance, through UK-based organisations, to operational partners in the disaster-affected countries?

Were there discernible advantages or disadvantages in terms of impact and efficiency in providing whole or partial budgetary support to the whole of a single agency's operations or financing specific component elements, eg equipment?

The role of ODA was quite different in the three countries and it is therefore difficult to generalise about efficiency and effectiveness in a wide range of different circumstances.

4.3.3 In Mozambique, ODA was, in essence, supporting through UK-based NGOs, a component of the overa 11 na t i ona 1 re 1 i ef programme, whi ch was organi sed pri mari 1 y to assist people affected by the war, many of them its displaced victims. The drought intensified food insecurity, particularly in the central and southern provinces, and ODA's additional food aid and funding for logistics was its response to the additional needs. Its assistance to those affected by war and drought was provi ded to a range of NGOs and to components of the Red Cross movement. The bulk of the food aid, prinCipally maize, was channelled through CARE, and during FY 1992/93 was distributed to beneficiaries before the start of the normal hungry season. ODA supplied 8% of the total amount of maize provided by the international community and was thus a significant 'actor'.

4.3.4 Assessment of the food aid component was made particularly difficult because of the poor monitoring and reporting and the lack of information on such specific aspects as: the condition of the population and the extent of their dependence on food aid; the degree to which the distributing agencies were able to provide pulses, beans and oils to complement the ODA-funded cereals; evidence of ODA-funded assistance's impact; and the effectiveness of food aid distributions undertaken by other agencies within the most seriously affected

32

Southern African Drought Impacts in Terms of ODA's Response

areas. Because of such difficulties it is only possible to indicate areas of concern rather than clear-cut conclusions.

4.3.5 Accountability was an area of concern. Monitoring and reporting were generally patchy and CARE was unable to prove receipt at the village level for a significant proportion of the food aid it distributed. Allowance does, however, need to be made for the insecurity and difficult operating environment in the country at the time.

4.3.6 ODA concentrated on the supply of maize as food aid rather than attempting to provide a balanced mix of cereals, legumes and oil. There is insufficient information to indicate whether the provision of a cereals-only ration heightened a serious nutritional risk for the recipient population. But where target populations are dependent on food aid over time, and where they do not have access to other sources of proteins, fats and micro-nutrients, a cereals-only ration will lead to a range of micronutrient deficiency diseases and increased risk of protein-energy malnutrition (PEM) in small children.

4.3.7 It is not clear whether the targeting of the ODA food aid was appropriate. Only 29% of the ODA-funded maize was distributed in the worst-affected Provinces which, according to the GoM's estimates, contained 63% of the drought-affected population. This geographical distribution reflected ODA's practice of inviting 'bids' for food aid from familiar, trusted NGO implementing partners, but also the fact that the activities of such partners were not concentrated in the areas most severely affected by the drought but were addressi ng needs primari ly resulting from the conflict. Such an allocation strategy had been introduced some years before and ECFAD was simply continuing an established practice. ECFAD personnel were aware that it would result in the bulk of the ODA assistance being distributed in areas other than those most severely affected by the drought, but considered this to be acceptable in that it would enable other donors to concentrate their assistance upon the worst-affected areas. The information available does not provide a sufficient basis to support or contradict this assumption. In retrospect, ODA might have explored the possibility of providing food aid to less familiar partners such as IFRC, ICRC and MSF which were active in the most severely affected areas, though this might have resulted in delays in the allocation process and greater risks in terms of dealing with agencies unfamiliar with its procedures and requirements.

4.3.8 The performance of individual programmes was of varied quality. On the positive side, substantial amounts of assistance were provided to populations seriously in need of relief. But other aspects of the overall programme, such as the acquisition of vehicles, could have been handled more efficiently by some of the agencies.

4.3.9 Overall, the international relief effort appears to have prevented widespread famine in Mozambique during 1992, although some areas experienced high rates of malnutrition, morbidity and child mortality. The food aid and logistics support provided by ODA will have played a positive role. In a difficult Situation, the overall programme seems to have been satisfactory rather than impressive.

4.3.10 In Lesotho ODA supported WFP in the local purchase of beans, provided funding to the Red Cross for the purchase of vehicles and made a substantial grant to Save the Children Fund to support logistical aspects of the Gol's food distribution operations. The last of these was a high priority area and appears to have been implemented effectively. The assessment information available i ndi cates that the food rel i ef aspect of the drought response was well co­ordinated and implemented by the NGO community.

33

Impacts in Terms of ODA's Response Southern African Drollght

4.3.11 In Zimbabwe ODA made grants totalling £0.97m to seven NGOs involved in programmes whi ch comp 1 emented the emergency food d is t ri but ions anJ 5 1JPp 1 ementa ry and school feeding programmes being undertaken by the GoZ. This funding was used to procure particular commodities, transport services or vehicles. Monitoring and reporting were patchy.

4.3.12 The total relief effort in Zimbabwe is considered to have prevented the drought shock creating a situation of serious food crisis with potentially localised famine conditions. Most of the activities supported by ODA appear to have been well-targeted, regionally, on the worst affected areas. Some programmes were impressive in their organisation and efficient in their delivery. Some specific questions, however, are raised concerning particular components of the ODA-funded grants. On the positive side, commodities were in some ~ases procured locally to save time in relief operations. But some commodity purchases and vehicle acquisition decisions may have been neither cost-effective nor particularly appropriate (see para 5.5.1(e) below). In other cases, ODA funding was used to pay for transport costs that should have been met by the original food donor.

4.3.13 Overall, the ODA-supported components appear to have been broadly successful in linking its response to differing circumstances, for example by funding local purchases as a complement to its own and other donors I food aid provision. The allocation of grants to logistics and distribution costs was more problematic. A significant number of the agencies supported in the three case­study countries used the funds to purchase vehicles either to transport food or to use to monitor operati ons. Despite a wi de range of ci rcumstances and experience, there are grounds for questioning the cost-effectiveness of these actions. Once obtained, it is not clear whether the vehicles were in every case effectively utilised.

4.3.14 Lessons:

Initial geographical targeting is perhaps the critically important decision. The evaluators were unable to determine whether, in Mozambique, the most appropriate geographic areas had been targeted for assistance.

As well as responding to requests from familiar partners, the possibilities of providing food aid to less familiar partners should be considered. Choices here depend critically on co-ordination with other donors.

Regarding sources of commodities, local purchases can provide flexibility and more timely responses. The appropriateness, however, of the commodities in question, as well as assessment of the local marketing implications, should be part of the assessment process prior to the making of a grant.

That a significant number of agencies more or less automatically turned to ODA to fund vehicle purchases or transport costs reflects on ODA's flexibility. That such activities can, when taken together, account for significant expenditures, suggests the need for a co-ordinated technical co-operati on input into future 1 arge rel i ef programmes to ensure that deCisions are made on an appropriate and cost-effective basis, and that eqUipment is effectively utilised.

4.4 RURAL WATER SECTOR COMPONENTS

4.4.1 Support by ODA for rural water programmes was substantial, the bul k of it accounted for by 5 large projects (2 each in Zimbabwe and Mozambique, and 1 in Lesotho). From several perspectives the performance of these actions was less

34

Southern African Drought Impacts in Terms of ODA's Response

than satisfactory and raises questions about the identification of needs and the appropriate ways of providing assistance in the water sector during a drought. The assessment of needs and the related targeting of priority assistance proved to be an area of variable performance.

4.4.2 Targeting: In Lesotho the activities chosen were not well suited to the emergency situation and geographical targeting bore 1 ittle relation to the distribution of people most at risk. In Zimbabwe the bulk of ODA's support was better targeted. It funded a geographical component of the National Emergency Programme, the Bikita District Project, and some of the actions of the NGO, Christian Aid. Both these actions were in the most severely affected south­eastern area. Similarly, in Mozambique activities appear to have been targeted on areas of high priority.

4.4.3 Appropriateness: In Lesotho the activities chosen were not well suited to the emergency situation and geographical targeting bore little relation to the distribution of people most at risk. The interventions supported ranged from transportation of water and rehabi 1 itating and improving water supply (by deepening wells) to providing new facilities by borehole and dam construction. Detailed consideration of these issues is undertaken in the case studies. The assessments for Lesotho and Zimbabwe, and the documentation available for Mozambique, confirm that only minor elements within two of the large projects, and another small grant to improve water supply for Mozambican refugees under threat of a cholera epidemic, were successful in addressing the water shortage problems during the drought period. The major part of the assistance did not yield benefits to the target populations until the 1992/93 rains had begun and the crisis had passed.

4.4.4 One reason for the lack of impact during the drought was the delay in submission and approval which is at least partly explained by the projects' complexity (see below, Chapter 5 Section 3). A second reason was the choice of activities. Transportation of water and repairs to facilities that might relieve the immediate needs of a community were not major aspects of emergency projects, except in Lesotho. Nor were actions that might quickly improve water supply, particularly well deepening, accorded particularly high priority, in contrast to some programmes of other donors in, for example, Zimbabwe. Instead, a large proportion of funding was directed to sinking new boreholes which, except on a small and highly selected basis, it was not possible to organise quickly enough to improve the water supply except in the longer term. Similarly, the construction of small dams by food-for-work combined immediate rel ief with the provision of structures that, if properly designed could improve potential water supplies. Consequently, the main achievement of the rural water component was to provide, after the drought, improved facilities for beneficiary communities that would probably make them less vulnerable in a future drought.

4.4.5 Where, as in Zimbabwe, local representative institutions were involved, sensitive to local needs and with knowledge of the local situation, activities were more likely to reflect local needs and priorities. In Lesotho the absence, or non-involvement, of such institutions was probably one of the causes of the ineffectiveness of the emergency water efforts.

4.4.6 The integration of water supply with health considerations was another, only partially effective, aspect of activities. Programmes appear to have been conceived primarily in engineering terms, to achieve supply objectives. In Lesotho there was no proper provision of head works or integration of water supply measures with complementary public health activities. In Zimbabwe, where contamination problems were arising with other well-deepening projects, the Bikita Project was re-designed after start-up to provide an enhanced component for the constructi on of head works. The rel ati vely better performance of the

35

Impacts in Terms of ODA's Response Southern African Drought

Zimbabwe project also illustrates how NGOs can have an important role in emergency water operations, where their involvement with local communities, commitment and low level of bureaucracy can come into play.

4.4.7 Lessons: If improving quantity and assuring potability of water supplies during a drought are priorities, then there are lessons to be learned from 1992/93 in terms of appropriate mitigation measures. Some other donor projects concentrated on rapidly implementable measures to improve supplies, such as well deepening and repairs, in order to have more impact during the drought. Slow start-up was another problem, suggesting that effective emergency water measures would benefit from more effective disaster preparedness. Institutional arrangements also make a considerable difference to effective response: the use of the NGO, Water Aid, as the implementing agency in a government-related programme, achieved more rapid implementation, some sensitivity to changing priorities, and more effective monitoring than in other projects. Emergency measures to increase water suppl ies must be properly integrated with health considerations. There is also the difficult decision-making problem under conditions of uncertainty during a drought, over the balance to be struck between measures to improve water supply almost immediately and measures that would have considerably greater benefit in the long-term or if the drought were to continue.

4.4.8 Finally, it is of key importance that ODA and those agencies it supports should monitor the water projects that it finances. There will be lessons to be learned in terms of appropriate mitigation measures and crisis response actions.

4.5 AGRICULTURAL REHABILITATION COMPONENTS

4.5.1 In responding to disasters which have destroyed or damaged crops and displaced rural populations it has become common to assist communities to recover their agricultural production and achieve a measure of self-sufficiency as quickly as possible. Typical measures have included the provision of seeds and other inputs, agricultural tools, access to draught power and finance for production efforts. NGOs in particular have included seed distribution components in their post-disaster assistance programmes undertaken, in areas or for groups, where subsistence agriculture predominates (Borton et al, 1988, 1989; Jones et al, 1991). Previous evaluations have had difficulty in assessing retrospectively the impact of such measures, particularly seed distributions. When combined with the frequent lack of objective assessment of the performance of agricultural recovery programmes by the agencies which actually undertook them, the information upon which to base a proper assessment is often unavailable.

4.5.2 A generally valid assumption underlying drought recovery measures is that farmers are unable to retain sufficient seed for planting and, that with the return of more favourable weather, recovery will be constrained. Marginal agricultural households are likely to be most severely affected by pressures to sell or consume their grain rather than retain it as seed. From a national viewpoint, the economic cost of commercial food imports justifies efforts to ensure that production is restored quickly. Speedy agricultural recovery also serves to limit the negative multiplier effects resulting from drought-related reductions in agricultural production, employment and incomes. These concerns justify major efforts to ensure the availability of seeds, inputs and agricultural credit, as well as price incentives for producers.

4.5.3 In areas affected by ongoing conflict, rural populations are often displaced and in such a situation agricultural recovery efforts require the provision, at the very least, of implements as well as seeds. Even where populations have not been displaced, their general drought-related impoverishment and the breakdown of normal marketing systems mean that there is often a need for

36

Southern African Drought Impacts in Terms of ODA's Resronse

agricultural recovery assistance which includes the provlslon of implements as well as seeds. Tools such as hoes, mattocks and machetes do need replacing from time to time and there is what appears to be a sound argument that the attempt to cultivate hardened, parched soil, increases the rate of wear. Most agricultural recovery programmes in Mozambique, therefore, included the provision of tools as well as seeds.

4.5.4 In Zimbabwe the lack of drought-resistant sorghum and millet led NGOs to organise a crash programme of imports and distribution. ODA supported Save the Children Fund in its seed distribution activities. SCF, on its own initiative, participated in a wider evaluation of sorghum and millet seed distribution by ICRISAT (Friis-Hansen and Rohrbach, 1993). That study, however, showed that the distributions were relatively late and that poor quality seed had been supplied. With the break in the drought and a massive increase in maize production (Table 2.1) these programmes were of little national economic significance. The targeting on poor, se1f-provisioning and severely affected communities was, however, socially beneficial.

4.5.5 In Mozambique ODA provided £l.Sm in support of the agricultural rehabilitation activities of NGOs, representing 17% of the expenditures in that country. The information available for the majority of actions in Mozambique was quite inadequate. Although most agencies claimed that yields were satisfactory, this cannot be substantiated because of the poor monitoring and reporting. Only one agency exp1 i ci t1y attempted to measure the yi e1 ds of the seeds it had distributed and only one appears to have succeeded in completing its distribution by the start of the rains in October 1992. None seems to have undertaken surveys of the recipients to obtain their views on the suitability of the seeds di stri buted or the time1 i ness of thei r di stri buti on. Much of the poor moni tori ng and reporting may be attributed to the difficulties of the operating environment but the fact that one agency was able to report fully suggests that others could have performed better in this regard. Overall, agricultural activities appear, on the basis of the available information for the actions reviewed, to have been a relatively inefficient part of the overall response.

Lessons

4.5.6 Taking into account the concerns of earlier evaluations, there appear to be three lessons for future agricultural rehabilitation, especially as operated by NGOs. First, such programmes and proposals should be prepared alongside those for relief measures because of long lead times and the likely poor performance of late distributions. Second, suitability and quality of seed are likely to be decisive considerations and NGOs may need technical support. Third, performance monitoring and impact assessment need to have higher priority.

4.6 HEALTH CARE COMPONENTS

4.6.1 The drought was initially perceived in terms of a regional food CrlS1S with serious economic (balance-of-payments) effects. Subsequently governments, the UN through the DESA process, and NGOs drew attenti on to its non-food dimensions, including the increased health risks, especially of diarrhoeal disorders, the nutritional stress amongst vulnerable groups and the intensified problems of obtaining potable water. In addition, the pressures on health institutions in Zimbabwe in both the public and the voluntary sectors to involve themselves in nutritional interventions, especially for pre-school children, temporarily drew capacity away from other emergency health functions and the normal programmes of immunisation, for example. Shortages of water and movements of population were associated with serious outbreaks of cholera in Mozambique and also in Zambia and Zimbabwe, along with reported increased incidence of dysentery, common shigellosis and common diarrhoea.

37

5.1 ORGANISATION OF THE RESPONSE WITHIN ODA AND THE POSTS

5.1.1 In responding to a region-wide emergency ODA utilised several aid modalities and budget lines. The complex programme of support involved particularly three ODA departments in London, as well as officials in post, usually those staff members of the High Commission or embassy with developmental responsibilities and, to a limited extent, the Development Division in Lilongwe, which provided advice. The response has been characterised as a phased process in paragraph 3.1.9. However, initial decisions on priorities in late February/early March 1992 critically determined the way the response was subsequently organi sed, in terms of provi di ng: government-to-government BoP support; food aid from the regular annual programme; financial and TC support, if necessary, to regional logistics; and tranches of emergency aid funds for relief and technical co-operation for which UK-based NGOs and, through the Posts, other organisations could make proposals.

5.1.2 The way the response was organised reflected different country situations, as illustrated by the three case studies and the phasing of ODA's response is described chronologically in Chapter 3. The roles played and the relationships between ODA, the governments of a ffected count ri es, i nterna t i ona 1 agenci es and NGOs are analysed here.

5.1.3 By reviewing all the actions undertaken during 1992/93 in the 3 case-study countries, 5 different patterns of response have been identified.

a. Di rect government-to-government BoP support: Approx i mate 1 y 40% of resources are accounted for by the sequential grants to Zambia and Zimbabwe. These followed Geographical Desk level assessments based upon initial indications from officials in post and affected governments of the sca 1 e of the resourci ng gap. ODA commi tments were announced directly to the recipient government. Subsequently, the Post was involved in ensuring that any administrative and reporting requirements that might delay disbursement were minimised.

b. Fundi ng emergency acti ons by government agenci es with bil atera 1 country funds: In Zi mbabwe, fo 11 owi ng an approach by GoZ for assistance with its National Rural Emergency Water Programme, the BHC, in consultation with London and with advice on feasibility by the Development Division, prepared a project for rapid implementation. When approved, an NGO, Water Aid, was brought in to administer the grant, to ensure its rapid implementation and provide TC inputs. ODA's assistance to Lesotho also involved direct financing of government action, through the latter's Emergency Rural Water Programme but without a British NGO involvement.

c. Responding to requests from international organisations: ODA was approached by, for example, WFP to contribute to its regional programme as well as through the Bri ti sh Red Cross Soci ety to contribute to the international appeal of the Federation of Red Cross Societies. These requests were handled respectively by ECFAD and EMAD in close consultation with the Geographical Desk. The contributions to the BRCS were earmarked for transport equipment as part of the national response by Federation members in affected countries.

d. International agency appeals and proposals at a country level: Agenci es such as UNICEF and WHO develop programmes for supporti ng national governments, for example with imported medical supplies and

41

Management and Policy Issues in the Response Southern African Drollght

vitamins likely to be required in programmes for combating malnutrition and drought-related disorders. For instance, in Zimbabwe, UNICEF and WHO initially approached officials in post and, after i nforma 1 agreement, proposa 1 s were passed from the country offices to the UN agencies' headquarters, which then made a formal request to ODA. The grant was channelled to headquarters which in turn devolved responsibility to the country office for organising supply, liaison and monitoring.

e. Grants to NGOs: Grants from EMAD funds in practice involved three different patterns.

i. Operational UK-based NGO proposals: A British NGO directly involved in emergency operations puts a proposal to EMAD or, for food aid, to ECFAD, to support its operations in the affected country. In practice, where NGOs are already on the ground, they may have informal contacts with the Post, especially if ODA is involved in joint-funding their development activities. A grant is made to the NGO headquarters which then channels assistance through its country office, which in return reports back via headquarters to ODA.

ii. NGO grants on behalf of 'local' partner organisations: The church-affiliated NGOs have traditionally prepared proposals or appeals at a country level which are then channelled through UK-based partners as requests to ODA for support. Wi th the growth of local secular NGOs, UK-based secular NGOs are increasingly acting in a similar capacity on behalf of their southern partners.

iii. Local NGO proposals: In this case the locally based NGO submits its proposal to Post officials. The proposal is then forwarded for scruti ny by EMAD, the CSAD Desk offi ci al and advi sers in London and, if approved, the grant is channelled through the Post which is responsible also for monitoring. This has, so far, been an exceptional practice in Southern Africa (for example the grant to LSM in Zimbabwe). As local NGOs become larger and more capable and as ODA strengthens its links with them, these relationships could become more common.

5.1.4 The above description of ODA institutional arrangements indicates a complex set of relationships that are not easily encompassed within any simple set of procedures for emergency assistance. In particular, the way the response was organised reflected differing country situations. In Mozambique, the continuing emergency led to the development of closer relationships with UK-based NGOs in-country. In addition, the NGOs had come to play a substantial role in providing relief and rehabilitation assistance within a relatively stable geographical division of responsibility. In this situation, responsibility for liaison with NGOs had been delegated, to a considerable extent, to the Post, with an important part of its role the supervision of emergency aid. In addition, EMAD funds were, in effect, delegated to the Desk, approximately £2m in recent years having been earmarked for continuation or initiation of emergency operations.

5.1.5 ODA accords the highest priority to the use of the annual UK bilateral food aid programme for emergency assistance and relief. In recent years, ECFAD has assigned part of the emergency allocation to Mozambique, using CARE (UK) as its channel. In allocating an additional £5m in response to the drought, ECFAD broadened the programme by soliciting proposals from other UK-based NGOs

42

Southern Africa Drought Management and Policy Issues in tile Response

operational in Mozambique, Action Aid, Christian Aid, Concern. Consequently, although proposals came from NGOs, ODA was, in effect, proactive. Only in small food aid grants to CARITAS and SCF, did the response represent wholly autonomous proposals generated locally by NGOs. These were met by transferring residual EMAD funds to ECFAD.

5.1.6 In Zimbabwe, ODA was already working closely with the World Bank and with other bilateral donors to support the economic reform programme. Given the large-scale imports required to sustain the government's own drought relief programme and normal food market channels, ODA, involving the Desk, ECFAD and the Post, developed its own proposal for providing additional BoP support. The Post, in consultation with the Desk and the GoZ, worked out the specifics of reporting arrangements to minimise disbursement delays. Some other grants also reflected informal consultation with government as well as with international agencies and NGOs at a country level. Otherwise, although there was local consultation on an informed basis between Post and NGOs, ODA was, in effect, reacting to proposals emanating from UK-based NGOs for their own operations in Zimbabwe and on behalf of their local partners.

5.1.7 Overall, the organisation of ODA's response was clearly influenced by existing aid relationships involving both governments and NGOs. The main differences arose as between countries where NGOs were al ready operationally involved and those where ODA had not previously provided NGO support. An issue for assessment of both the effi ci ency and effecti veness of ODA' s response is whether there were observable differences in performance between these different circumstances.

5.2 ODA's RESPONSE STRATEGY

5.2.1 In responding to a large and complex regional emergency, ODA quickly formulated a strategy which shaped its subsequent actions. Thus, it did not simply react to appeals from affected governments, international agencies and NGOs. The strategy, as already described in more detail in Chapter 3, involved a phased approach and employed a range of instruments in seeking to make responses appropri ate to di fferi ng country ci rcumstances. Thi s expl i ci tly phased response involved providing a first tranche of assistance as quickly as possible and encouraging others to do the same, whilst awaiting information that would indicate the scale of the eventual overall response. In acting in this way, ODA appears to have learned two lessons from previous emergencies. International agencies, particularly the WFP, had moved towards favouring a phased res~onse to disasters of a potentially large but initially uncertain magnitude.' Second, ODA recognised the economic aspect of a drought shock, and it had some previous experience of responding to drought with BoP support for Kenya during the 1984/85 drought (Borton et al, 1988).

5.2.2 The content of the first tranche of assistance reflected ODA's decisions on both priorities and targets. The subsequent tranches provided blocks of funding which left specific priorities to be determined by NGOs in making their proposals. An overview of grants made from EMAD funds, in particular in the three case-study countries, raises a number of issues concerning this 'reactive' approach. ODA informally indicated to NGOs at both a country level in, for example, Mozambique and Zimbabwe, and in London, that agencies could bid for ODA support withi n an overall budgetary ceil i ng. As ODA had in effect already determined how the greater part of its response would be used, this was in effect a supplementary programme complementary to the principal types of support. This NGO programme therefore effectively delegated primary responsibility for the

.. The rationale for the establishment 01 the Immediate Response Account in 1991 by WFP was to increase its capacity to make an initial response prior to a more detailed

one using the resources of the IEFR.

43

Management and Policy Issues in the Response Southern African Drought

determination of country, sectoral and intra-country regional priorities to operational NGOs or their UK-based partners.

5.2.3 A review of NGO actions indicates their characteristic response to this funding availability. Aware of the overall amounts available, the major UK-based NGOs put part of their overall funding requirement to ODA for specific country actions. The proposals reflected their knowledge of various donors and their preferences, for example the likelihood that food aid would be available trom the EC through EuronAid. ODA is known to be relatively flexible in providing financial assistance; proposals therefore typically included local purchases of food, acquisition of transport and other equipment, and internal transport and other operational costs. Where the operational agency is a local partner, the UK or EC-based NGOs are, in effect, acting as brokers. Proposals are directed, often after some informal consultation, according to their knowledge of donor preferences and the levels of funding likely to be forthcoming.

5.2.4 In responding to NGO requests, ODA was often funnell ing resources into quite complex international networks of agencies. For example, proposals made by CAFOD for Mozambique involved grants which were channelled to CARITAS Germany, and then on to CARITAS in Mozambique. In Zimbabwe, OXFAM and Christian Aid as well as OXFAM Canada and the Dutch agency NOVIB, were all supporting ORAP's relief and other operations, with some duplication of requests as a result of a lack of co-ordination within the group.

5.2.5 The conventional idea of relief and grants made by ODA is represented more by the directly operational agencies such as Concern and SCF. An increasing proportion, however, of relief assistance is being channelled indirectly through partners. This has several implications. First, it is difficult for ODA to assess requests where it is being asked to fund part of a larger programme financed by several donors. Second, there is a possibility of multiple funding of the same activity. Third, there is some risk of dilution of accountability, especially where the involvement of local partners constitutes an additional organisational 'tier'. In fact, in some cases, the local partner may be unaware of the nature of ODA' s i nvol vement or its requi rements regardi ng either the submission of proposals or the monitoring of grants.

5.2.6 A further issue is whether the allocation of a tranche of resources with an invitation to bid is an efficient mechanism for emergency relief. NGOs, such as the members of the UK Disasters Emergency Committee (DEC), which are closely acquainted with ODA and its procedures, are likely to take a view of how much they can obtain from a particular tranche of resources and the size of proposals appears broadly to reflect this. Within ODA there was also an expectation that funds were 1 ikely to be spread across the major UK-based NGOs active in responding to the drought emergency. In fact the cost of proposals submitted did not significantly exceed the sums initially allocated and indicates, either by accident or by a process of informal rationing, such a spreading of grants across agencies and countries in the region. The process did not appear to be in any sense competitive.

5.2.7 A possible implication of this practice, established over a sequence of emergencies, is that the quality of some proposals, and the basis on which these are submitted, is less than satisfactory. In one case an NGO considered it reasonable to submit no more than a two-page faxed proposal for a grant of £500,000. In another, a proposal for supporting agricultural rehabilitation (not a life-saving activity) was accompanied by a request for it to be approved the same day.

5.2.8 Where proposals come through intermediaries these do not necessarily meet ODA's requirements. These inadequacies do not of course necessarily imply that

44

Southern Africa Drought Management and Policy Issues in the Response

the activities to be undertaken are in any way unsatisfactory. In assessing proposals submitted to EMAD, ODA regularly seeks the opinion of Desk, Post and, as appropriate, advisers; its concerns include targeting aspects in relation to region and beneficiary populations. Because agencies cannot necessarily be assumed to have co-ordinated their actions, an issue is to avoid funding overlapping proposals. With information from the Post and other sources, ODA also attempts to form a sense of targeting priorities. Accountability for the use of public funds also requires that proposals provide an indication of likely uses. Cases were identified of proposals where delays occurred because an activity apparently duplicated another that ODA was planning to fund. In another case, a proposal for retrospective funding was inaccurate at the time it was submitted, leading to extensive consultation about how the initially unaccounted­for funds would be utilised. An important implication is that ODA's efficiency in responding cannot be considered in isolation from the efficiency of those who are seeking funding and implementing activities.

5.3 EFFICIENCY AND TIMELINESS OF RESPONSE

5.3.1 The efficiency aspect of ODA's response has been examined in various ways. First, in responding to an emergency, timeliness is an important issue. Even though drought can be characterised as a 'slow-onset' disaster, the risks of discontinuities in the food pipeline, and the health and other problems associated with deterioration in the quantity and quality of water, imply some urgency. That sense of urgency is particularly the case in the provision of assistance to landlocked countries where, inevitably, there will be a period of several months before imported food is delivered and distributed. In responding to the Southern African drought, ODA initially moved quickly. Its initial tranche of support was committed on 4 March only a month after Zambia declared an emergency and before the Government of Zimbabwe made its official declaration and request for international support.

5.3.2 In addition, it was also considered worthwhile to look at the time from start to completion of operations organised on an emergency basis. The results of this analysis are presented in Tables 5.1 and 5.2, taking 1 April, the day after ODA indicated that it would consider NGO proposals, as the start of the emergency. This analysis shows that there were quite substantial time-lags in the process of preparing and submitting proposals to ODA of, for example, 14 weeks on average for non-food aid grants for Mozambique and Zimbabwe. ODA's approval time was shorter (7 weeks for non-food aid actions). The duration of operations was highly variable.

45

Management and Policy Issues in the Response Southern African Drought

Table 5.1 Analysis of timeliness of ODA's response to 1991/92 drought in three case-study countries

Rn:ipi~nl Stl'lor No "f ;,,;!i,'n~

- Non [,,,,ll ai<l 17

Zimhabw~ 17

L"'snt!w

SOE(a) 10 Ik"l'ipl(l[

Ih'n'ipl"r Ile'l"l'."l I"

H~qllc.\t(h) AI'1'''''',11(c')

Il

" " OR

SOE In Hen'illl,'( l),n,,!jnn "I

SOO Ik'l"""\ I" 500 In!,'r\'<'ntinn

(d) (c)

" ''I J9 J.I 21 " 27 " 21

21 Il 47

Notes: a. Start of Emergency (SOE), taken at 01.04.92 following ODA meeting with UK NGOs. Analysis excludes actions that were initiated before this date.

b. c. d. e.

Date of arrival of request at ODA. Date of official approval of funding by ODA. Start of Operation (SOO) - see Table 5.2 note (d). From SOO to end of intervention (eg last distribution of food aid commodities etc).

Table 5.2 Analysis of timeliness of ODA's responses to the 1991/92 Drought in Lesotho, Mozambique and Zimbabwe by sector

Hccipicnl SC\:l11f ~" pr aC'li(\n~

Agri<:!d!urc(1) <)

F""d ~;J <}

"""d Di~l!ibu(i"Il(!!) 1(' Heallh J Waler

Ollicr(h)

SO[(,,) to lin'cipl(\r R<'c'(ipt or R(qt!<'~II(l

HC''1"c-''!(h) '\l'l'r<,,":1I(<:)

22 12 Il 33 11

"

SDI, I" RC<"d!,! pr l)"r.11;"11 (lr soo Rcqllc'st In SOO Inl~rvtll!i"f\

(d) (c)

.~~ ", " '"

,., 27 1(, ,.' .17 " .' " <, ·17

26 17 >C,

Notes: a. Start of Emergency (SOE), taken at 01.04.92 following donor meeting and declaration of national emergencies. Analysis excludes actions that were initiated before this date.

46

b. c. d.

Date of arrival of request at ODA. Date of official approval of funding by ODA. Start of Operation (SOO) , defined as:

Food aid - date of ~irst distribution to beneficiaries Seeds and tools date of first distribution to beneficiaries Transport - date of delivery of trucks or payment for spares Water date of commencement of dri 11 i ng or other intervention Health - date of first contribution of drugs or other intervention Where exact date of SOE is unknown approximations to the nearest month have been used.

Southern Africa Drought

e.

f. g.

h.

Management and Policy Issues in Iha Response

From SOO to end of intervention (eg last distribution of food aid etc). Agricultural rehabilitation, seeds and tools distribution. Feeding programmes, transport and logistical support to feeding programmes. Multisectoral interventions.

5.3.3 A range of specific contributory factors accounts for these time-lags. The situation in Mozambique is in some respects distinct because the drought involved additional relief rather than the organisation of new activities. Consequently, there was perhaps less sense of pressure. However, in Zimbabwe, where the drought began to impact severely by April, the average time lapse of 27 weeks to start of operations indicates some regrettable delays. In Lesotho, the difficulties of NGOs working out an effective drought response strategy with what many regarded as a slow moving government, delayed the formulation of specific proposals.

5.3.4 The analysis indicates that the process of formulating and submitting requests is itself a source of substantial delay. Directly operational agencies appeared to act more quickly. Where the process involves preparation by a local agency and the indirect channelling through possibly an international partner or even group of partners, further delays may be involved. A scrutiny of individual grant actions indicates some of the particular causes of delay. Insufficient information on whatever was intended often necessitated a process of discussion involving officials in post. More complex proposals, for example for emergency water supply, were correctly subjected to more detailed scrutiny than those for relatively simple relief food operations or purchases of equipment. Differences in in-country knowl edge of emergency procedures may be another factor. For example, in Mozambique, where emergency operations were a normal part of officials' relations with the GoM and NGOs, food aid grants were approved relatively quickly. In contrast, in Zimbabwe, where emergency operations were a new activity, both the Post and even local offices of international organisations were not necessarily familiar with procedures and, as in the case of a health sector grant to UNICEF, some delays occurred.

5.3.5 The period from request to approval appeared in most cases to be the only aspect of an activity where ODA's actions contributed to determining the start of operations. In almost all cases, funds once approved appear to have been released very quickly. Managers in partner agencies invariably commented favourably on ODA's good practice in this regard, and also on its flexibility as regards the actual use of funds released at the outset of an emergency operation.

5.3.6 In sum, the extent to which an action can be considered to be timely depends on the type of activity. For food distributions the objective should be to sustain supplies to the target population beginning as soon as possible after the (failed) harvest. Many self-provisioning rural populations experience a 'hungry season' in the months immediately prior to a harvest. The effectiveness of food aid is reduced if it is not provided in time for the start of the hungry season. It may even have a negative impact in terms of reducing the price obtainable by farmers for their produce and possibly also resulting in their modifying their cultivation for the following season. It is vital that seeds and tools are distributed to farmers in sufficient time for planting and achieving an optimum yield. In the case of emergency water projects undertaken as a response to drought, effectiveness requires that they increase the availability of water to the target population as soon as possible. Projects which do not achieve this before the next rains begin cannot be considered effective in relation to the needs created by the drought, though they may bring longer-term benefits.

47

Management and Policy Issues in the Response Southern African Drought

5.3.7 Of the 8 separate agricultural recovery actions evaluated - 6 in Mozambique and 2 in Zimbabwe - only one appears to have completed distribution by the start of the rains in October 1992 ie, that undertaken by World Vision in Tete Province, Mozambique. Most of the others appear to have been undertaken during the period November 1992 to February 1993.

5.3.8 In the water sector, 6 separate actions were supported by OOA, one of them involving the rapid provision of water to Mozambican refugees. Thus 5 actions involving a total expenditure of £2.3m can be considered to have been responses where an important objective was to relieve the problems of access to water resulting from, or accentuated by, the drought. Only 2 of the actions, the only ones to have included a tankering component, can claim to have relieved water stress in advance of the October rains. None of the components involving well­deepening, borehole drilling and equipping of wells/boreholes with handpumps had an impact prior to the start of the rains. That such a large proportion of the measures supposedly designed to relieve the water stress created by the drought failed to have an impact before the rains relieved that stress is remarkable. In part this may be due to the presentation of ongoing rural water programmes, in what are generally drought-prone areas, as drought responses in a situation where funding may be more readily available for 'emergency' than for 'development' activities.

5.3.9 OOA's approach to providing BoP support for drought-related imports was generally the subject of favourable comparison with other donor procedures (see Chapter 4.2).

5.4 COST-EFFECTIVENESS

5.4.1 The complexity of actions supported partially or wholly by OOA made it difficult to analyse the cost-effectiveness of responses in any systematic way. A particular difficulty in examining emergency operations is that many choices were made to ensure timely action and this frequently involved acceptance of higher costs. A further difficulty lay in establishing benchmarks for measuring cost-effectiveness, for example in Mozambique where local transport arrangements in a war-affected country were subject to change in different areas. Confronted with these di ffi culti es, the approach adopted was to look for evi dence that individual operations were being handled with regard to value for money. This impl ied seeking to ensure that ODA, its grantees and their local partners explicitly looked for low-cost solutions to specific requirements wherever it was realistic to do so in terms of time pressures and specifications.

5.4.2 The evidence from the three case-study countries indicates that most grantees sought to be cost-conscious in their purchasing activities. Where, however, there were choices between different modes of operation, particularly as regards transport and logistics, there are unanswered questions concerning the cost-effectiveness of some actions. These general refl ections are ill ustrated by a review of cost-effectiveness on the basis of the various components:

48

a. Financial Aid: ODA allowed the Zimbabwe Government to invoice the costs of oil imports already made against financial grants, and so this did not involve any direct examination of the tendering and procurement procedures for drought-related imports. As, however, food and agri cul tural import purchases were bei ng covered by lOA credits, under normal World Bank procedures, a framework was provided for a reasonable presumption that these resources were being used cost-effectively.

b. Food and Log; st i cs: CARE, in agreement with ODA, tendered for ye 11 ow maize on the international market and purchased from the lowest cost

Southern Africa Drouoht Management and Policy Issues in the Response

source, Argentina. The lack of comparability between the various internal transport and handling arrangements, in war-affected Mozambique, precluded an analysis of ITSH costs.

c. Some local food purchases for relief activities were made in Zimbabwe using EMAD grants. Dried skimmed milk was purchased locally and was available for immediate delivery at a lower cost than if it had been procured in Europe by the European Commission, the potential alternative source of supply of dairy food aid. Taking into account the costs of transport from Europe, this was, therefore, a highly cost-effective solution. NGOs also made small purchases of elephant meat and other products in Zimbabwe. Local purchases of maize meal in particular were well below the costs of imports, and in that sense represented value for money. Nor were these operations on a scale that might have had implications for local markets.

d. There is insufficient information to determine whether or not local purchases in Mozambique were cost-effective. A few actions raise questions of what is acceptable practice. The local purchase of yellow maize for relief distribution apparently made sense as the least-cost option in some local ities. However, this involved the re­acquisition of food that had originally been provided as food aid and then 'monetised' by the government. If justified by local conditions, it nevertheless implied the double counting of food aid in terms of overall assistance to Mozambique.

e. Vehicle purchase: agencies were often balancing the need for rapid acquisition, cost and specification differences. Some of these choices appeared retrospectively, for instance in Zimbabwe, to have been unsound. A wider issue of cost-effectiveness is whether the preference for own operation rather than use of commercial contractors was always fi nanci all y sound. Vari ed experi ences on purchasing and the problems of some agencies in covering running and maintenance cost suggest that logistics and vehicle purchase should be an area requiring technical assistance in a major emergency such as that in 1992.

f. Rural water supply: The investigations in Lesotho and Zimbabwe indicated that particular activities had been organised in a fairly cost-effective manner. For example, a comparison of the costs of well drilling operations did not reveal any conspicuous case of cost­inefficiency. The unresolved issue concerns the balance of activities chosen in terms of the cost of achieving improvements in water supply during the drought and any longer term benefits of improving water supply. Where benefits of the latter form are expected, then an acceleration of normal programmes for improvement supply is more likely to be cost-effective than an improvised cash programme such that in Lesotho.

g. Health component: The acquisition of drugs appeared to have involved a successful balancing of considerations of cost and timeliness, for example in the choice of shipping or air transport for different categories of drugs.

5.4.3 Some activities were unquestionably cost-effective in meeting narrowly defined requirements. Others were, however, less evidently cost-effective, despite the broad concern of agencies to get value for money in their specific purchasing. Transport and logistics and local purchasing of food relief are areas in which there is need for clearer guidelines. Without a clear

49

Management and Policy Issues in the Response Southern African Drought

understanding of the objectives of the emergency role of water programmes, the cost-effectiveness of different courses of action is difficult to explore.

5.5 EFFECTIVENESS OF CO-ORDINATION ARRANGEMENTS

5.5.1 The international community's co-operation with the governments of the affected countries is by general agreement a success story. The food import and logistics aspect of the drought response was unprecedented in recent experience. The exercise involved WFP working with SA DC and the transport systems of South Africa and SA DC countries. The establ ishment and operation of the Logistics Advisory Centre as well as the various corridor groups for managing overland transport are now well documented as critical elements in the co-ordination. The arrangement was initially concerned with mainly commercial imports, as food aid, with the notable exception of that from the US, took some time to begin arriving in large quantities (SADC, 1993a: 70). In fact, the co-ordinated response on the food aid side was somewhat late and not wholly satisfactory, with a median a rri va 1 date of November or December 1992. It effect i ve 1 y began with the 25 February meeting in Brussels. That meeting, critical to internationally co­ordinated response on the donor side, was in a sense fortuitous. It happened to provide an appropriate context but there is no provision for such meetings on a regular, for example, half-yearly basis.17 At that time, furthermore, there was no parallel arrangement for formal or informal consultation amongst donors concerned with disaster relief apart from food aid. The relatively stronger element of donor co-ordination in the area of food aid also perhaps explains the somewhat unbalanced overall response to the Southern African drought.

5.5.2 Although there was international consultation, ODA correctly assessed for itself that the most seriously affected countries would be confronted with large financing problems for commercial imports and established its priorities on that basis.

5.5.3 The UN Department of Humanitarian Affairs (DHA) took the responsibility, in consultation with SADC, of co-ordinating the overall response. This process was relatively slow, resulting in the consolidated DESA appeal at the end of May 1992. On the food aid side, the proposals were already too late; governments had already been obliged to organise commercial imports, and most of the donors had begun to make commitments. When DESA appeals were compared with what was known about the proposed activities of the World Bank and bilateral agencies, including their co-operation with NGOs and the governments of affected countries, the appeal appears not to have been well co-ordinated and even to have involved duplication. Retrospectively, it is difficult therefore to assess the overall response of the international community using the DESA as a yardstick. Governments were making large-scale commercial imports, but with some additional assistance, for example from ODA as well as the World Bank. Subsequent DESA attempts, to which ODA gave partial support, are also difficult to assess. The food aid requirements of the region as consolidated in the DESA appear to have been largely met. However, with the end of the drought and relatively favourable growing conditions in a number of affected countries in 1992/93, the FAO/GIEWS, for example, was already indicating possible excessive levels of imports in relation to requirements during 1993/94.

5.5.4 The role of DESA in relation to the non-food requirements of drought­affected countries is even more difficult to assess. The original appeal included only the proposals of UN agencies in co-operation with the governments of affected countries. Subsequently, DESA monitoring appears to have only partially included the response of bilateral donors, including activities

17 In lact, the practice 01 hQlding these meetings appears to have lapsed.

50

Southern Africa Drought Management and POlicy Issues in the Response

financed by ODA, and not all World Bank activities were taken into account. The widely reported conclusion of DESA that the international response to the non­food requi rements of affected countri es was inadequate seems di ffi cul t to substantiate (see Section 3 of Chapter 2 esp. paras 2.3.8 & 16).

5.5.5 The World Bank, once it too had recognised the potentially serious scale of the drought in February 1992, played a major role in focusing co-ordinated attention on the economic dimension at a country level through its consortia and consultative group arrangements. It also made an unprecedentedly large response, including $345m of additional and redirected credits, but encountered disbursement difficulties. The IMF also made an important contribution by drawing the attention of Board members to the large resource needs created by food and other imports. These various actions contributed to the consistently high priority accorded by virtually the entire international community to working with the governments of affected countries to minimise the human and economic effects of the drought. ODA, as one of the more important bilateral donors in the region, could therefore act in reasonable confidence that it was contributing assistance that would be complemented by food aid and financial support by the rest of the donor community.

5.5.6 Co-ordination at a country level is considered in more detail in the case studies.

5.6 EFFECTIVENESS OF MONITORING AND REPORTING ARRANGEMENTS

5.6.1 NGOs receiving grants using EMAD and ECFAD funds are required to report on operations within six months of commencement of activity, and to make a final report within six months of completion. The emphasis is on accountability rather than impact assessment. The onus is therefore on the grantee to determine to what level it chooses to provide information to ODA as the funding agency. Where, however, bilateral aid funds are involved within a project framework, the monitoring and reporting requirements are more precise and detailed.

5.6.2 It is no surprise, therefore, that reporting practice observed for grants in the case-study countries was variable. Most grantees appeared to have met minimal reporting requirements in terms of some statement of activities and expenditure according to the preliminary reporting requirement, but many had not made final reports at the time of the evaluation, even though they were due. In part, the minimal reporting appears to reflect an NGO view that ODA is primarily concerned with financial accountability and fails to provide any feedback. Accordingly, there was no incentive to do more than meet minimal requirements.

5.6.3 The monitoring of grant actions by the Posts was also highly variable. In cases where the Post had been involved in initiating grants and had.a monitoring responsibility, for example with Water Aid and LSM in Zimbabwe, or where problems arose with proposals, a closer interest was taken. However, in the case of grants made with EMAD funds to UK-based NGOs on behalf of local partners, the Post was not required to spend time monitoring grants channelled indirectly in this way, and appeared to have little incentive to do so. The effectiveness in monitoring grants also depended on the strength of staffing. For exampl e, the BHe in Zimbabwe was abl e to draw on the servi ces of the Development Division in Lilongwe for advice in monitoring the Bikita water project supported with bi 1 ateral project funds. The grant to support the government's Rural Emergency Water Programme in Lesotho appears not to have been the subject of any technical monitoring.

5.6.4 Four issues arise in reviewing the effectiveness of monitoring and reporti ng performance: accountabi 1 ity for publ i c resources; 'techni ca 1 ' monitori ng and opportuniti es for improvi ng performance; impact assessment -

51

Man~genwnt and Policy Issues in the Response Soutliern African Drought

better understanding of the effectiveness of resource use, offering lessons for the future; and accountability to the intended beneficiaries.

5.6.5 Currently, monitoring and reporting are largely concerned with accountability for public resources. In the case of intergovernmental actions the normal procedures appear to be satisfied. Where grants are made to NGOs, reporting depends almost enti rely on the practi ces of the agency concerned. Where ODA is 'the back donor' for funding of NGO partner activities, reporting and monitoring depend on the conditions imposed by UK NGOs. Under the pressure of an emergency, organisations are perhaps reluctant to commit scarce management resources to what are seen as routine rather than operational activities. In cases where ODA is only one of several sources of support, there are inevitably problems of attribution. It would be more appropriate therefore to require reporting and assessment of the whole activity, perhaps in conjunction with other supporting agencies.

5.6.6 'Technical' monitoring, linked in some cases to assistance in problem­solving, appeared more likely to occur where the Post had been involved in initiating the project or had a specified monitoring function. Where ODA was providing support indirectly and even the UK NGO had no local presence, the possibilities for such monitoring are limited, as they are in a situation of internal conflict as in Mozambique, though even so there still appeared to be considerable room for improvement in agencies' monitoring and reporting activities. The evaluation's ability to assess the effectiveness of many actions was severely limited by the patchiness of such 'technical' monitoring, even in relation to basic sorts of information.

5.6.7 Impact assessment, which may often build on technical monitoring, can provide valuable feedback on the effectiveness of emergency operations. A few organisations succeeded in assessing the performance of their activities. For exampl e, SCF attempted to moni tor the performance of both its suppl ementary feeding and agricultural rehabilitation activities in Zimbabwe in considerable detail; its reports were amongst the few informative impact assessments undertaken during the drought. The World Vision assessment of seed distribution in Mozambique indicates that this was possible even in a confl ict-affected country. Further efforts in this direction by the agencies supported by ODA would therefore be worthwhile. For example, technical monitoring of the water projects in Zimbabwe provides useful information on the relative effectiveness of different types of intervention.

5.6.S The views of beneficiaries can be a valuable tool for assessing how effective and appropriate assistance has been. Though most agencies will have had frequent contact with the beneficiary population and their representatives, there were virtually no attempts by agencies to seek the views of the beneficiaries in a systematic way, or to sample beneficiary populations, to assess the effect of activities.

5.6.9 The normal monitoring requirements for ODA emergency grants do not include any form of detailed technical monitoring, or assessment either of impact or the recipient population's views. Some donors do build such requirements into their grants and include funding for these purposes. This evaluation has in some cases had to rely on information from these assessments, for example, the ICRISAT Seeds Distribution (Friis-Hansen and Rohrbach, 1993) and Africare Regional Water Programmes (Mason and Lebane, 1993) for USAID. Otherwi se the assessment of impact has had to depend on information available because some grantees gave higher priority to being informed about their own activities (eg SCF in Zimbabwe) and on the specific investigations, undertaken as part of the evaluation, into water projects in Lesotho and Zimbabwe.

52

Southern Africa Drought Management and Policy Issues in the Response

5.6.10 After a peri od of consul at ion wi th NGOs and bi 1 atera 1 donors, ODA issued in May 1994 "Humanitarian Assistance: Guidelines on Project Proposals and Project Reports". This sets out the detailed requirements for project proposals and also for completion reports which should be submitted within three months of the completion of a project. This should help ensure that information is gathered on implementation and effectiveness.

5.6.11 That so few agencies attempted to assess their effectiveness by examining the impact of their activities considerably limited the evaluation. These problems are of concern not only to evaluators; they also prevent operational relief agencies, as well as donors such as ODA, from developing a sense of 'what works' and what 'doesn't work' so as to improve their performance in subsequent emergencies. According low priority to monitoring and reporting is in the end a false economy in relation to the magnitude of resources otherwise allocated (WFP, 1994, Para 10).

53

Southern African Drought Conclusions

6.1 The conclusions from the evaluation are organised according to the seven items posed in the Terms of Reference for the study in Chapter 1. In addition, the effectiveness of monitoring and reporting is also considered. Finally, the criteria in the Evaluation Success Rating Table for summarising the performance of an emergency operation are examined. All these more specific conclusions concerning ODA's response and the problems of assessing that response should, however, be placed in context.

6.2 The response of the affected countries and the international community to the 1991/92 drought is widely regarded as a success story, and this evaluation broadly endorses that conclusion. ODA itself performed well in responding early and with a coherent strategy. Both the international community more generally and ODA appear to have succeeded in incorporating many lessons in their emergency practice since the previous regional crisis in Africa in 1984/85.

6.3 Nevertheless, a closer scrutiny indicates areas of less satisfactory performance. There is scope for improvement in the way ODA, NGOs and the international community prepare for and respond to drought shocks. The need for this should be stressed, given the likelihood that there will be another major drought-related crisis in Africa, or another major natural disaster elsewhere, before very long.

6.4 ODA's general approach: ODA performed well in terms of identifying the worst affected countries and devising its overall response strategy at an early stage. That performance resulted from the effectiveness of CSAD, in consultation with the officials in post, in monitoring the situation in the major countries of the region, and the maintenance by ECFAD of a watching brief on food security and food aid.

6.5 The principal weaknesses were found in terms of the way that ODA organised its response at a country level where strategies were less clearly developed. In particular:

a. It did not have a clear strategy for its actions in Mozambique and consequent 1 y allowed NGOs, through whi ch vi rtua 11 y all its ass is tance was channelled, to set the agenda. Most of these agencies preferred to direct additional drought-related resources to their existing relief programmes which were not in the areas most severely affected by the drought. Whilst not one of the largest donors, ODA, as a significant donor, should have taken a more proactive role. That would probably have required reinforcing the Post with someone with relief experience.

b. It failed to question rigorously the content of many proposals submitted by agencies, some of which were uninformative and poorly prepared. Thi s suggests the need for a strengthening of the technical capacity within ODA for reviewing proposals, including those for food aid and emergency grants. The poor quality of some proposals suggests the need for more guidance to agencies submitting proposals. (As indicated in paragraph 5.6.10 this was provided in May 1994).

c. Monitoring capacity in-country was too limited to allow a necessary minimum number of monitoring visits and follow-up on agency reporting. Consequently, for example in Zimbabwe, the officials in

55

Conclusions Southefn African Drought

post tended to concentrate thei r attenti on on the few projects initiated with their direct involvement. In particular, most grants made to UK-based organisations, sometimes for partner agencies in­country, were not satisfactorily monitored (see below).

6.6 The speed of response: The balance-of-payments support was provided speedily and, despite being a new form of emergency response, smoothly.

6.7 The bulk of the food aid provided to Mozambique during FY 1992/93 was distributed before the start of the normal 'hungry season' - 90% of it by the end of 1992. Thi s compares favourably with food ai d provi ded by some other donors. The performi1nce, however, of agricultural recovery and emergency water programmes was less impressive. The major part of the assistance did not yield benefits to the target population until after the 1992-93 rains had begun and the crisis had passed.

6.8 Analysis of the proposals submitted to ODA indicates that many NGOs were slow in submitting proposals for non-food aid assistance and consequently their actions often started late. In the case of some NGOs submitting proposals on behalf of partner organisations in the affected countries, the poor quality of the proposals resulted in further delay. ODA's performance in terms of the time taken to process and approve the proposals was satisfactory in most cases.

6.9 The value of the major elements and value for money: The main categories of expenditure are summarised in Table 6.1. Overall, ODA appears to have obtained value for money. Local purchases of food for relief, where direct comparison could be made, compared favourably with costs of food aid imports by other donors, eg skimmed milk powder in Zimbabwe. Because of the relatively late implementation of many of the agricultural rehabilitation projects, many agencies had to pay a premium which could perhaps have been avoided had they arranged purchases earlier. In truck purchases, some agencies appear to have opted for the better deals available from South African suppliers. This may not, however, always have led to the purchase of the most suitable models. The procurement and shipping of drugs and other medical suppl ies by UNICEF appears to have been generally cost-effective. It is important to stress that the issue of cost­effectiveness cannot be considered separately from that of timeliness.

Table 6.1 ODA Bilateral Response to the Southern African Drought by Recipient Sector

Re(/pirnf Sector

II<IIIHKC of ]',1ymcnfs

Wnlcr lIenlth Al!rkuUur~1 Hehahilifal;<>n FUlld Aid nnd I)i~trihll(i"n

Mulli SC\'lur

£ million

20.0

.'.1 l.l 4-5

17.0

3.6

49.2

':'~ of/otal

40.7

6.2 2.2 9.1

34.6

7.3

100.00

6.10 Impact of the major components: There were difficulties in determining the impacts of most actions supported by ODA. Impacts can only be inferred indirectly by the way in which affected groups, regions and countries recovered without evidence of intensified problems of economic distress, eg migration and increases in morbidity and mortality. A further problem is that the interventions supported by ODA represented only part of the overall drought response, even in one district or province.

56

SO!lthern African Drought Conclusions

6.11 The other reason why it is difficult to determine impacts is because only a few of the operational agencies involved attempted to measure impact. In Mozambique the security situation meant that it was especially hazardous for agencies to monitor distribution or to carry out surveys of impact. The bulk of ODA food was channelled through CARE. There was, however, no proof of delivery to beneficiaries for a substantial proportion of it, because of a change in transportation arrangements and a lapse in monitoring by the agency. That World Vision did at least attempt to assess the impact of a seed distribution in Mozambique suggests that others might have followed suit. Given the scale of the assistance provided and its importance to the intended beneficiaries, agencies shoul d have stri ven to assess the impact and effecti veness of thei r rel i ef actions more frequently at least on a selective basis.

6.12 An assessment of the major benefits delivered in relation to costs: Inevitably, given the lack of information on impact assessment, assessment of benefits in relation to costs is at best impressionistic.

6.13 The benefits of financial aid for balance-of-payments support are inherently difficult to assess. However, grants to Zimbabwe and Zambia were extremely well-timed and fast acting. This assistance contributed to the success of the government's efforts to avoid the breakdown of the normal food marketing system and the serious economic and pol itical consequences that would have followed. The additional assistance also reduced the government's financing requirement for drought response, thereby reducing the negative impacts of the drought on its medium-term economic reform strategy.

6.14 Though the impact of the food aid in Mozambique could not be assessed, most of it was distributed during the period when it was needed.

6.15 The lack of information on the yields obtained from most of the seeds distribution programmes means that it is not possible to comment on whether the benefits outweighed the costs. Had the distributions been more timely the likelihood of this would have been enhanced.

6.16 The probl ems of timel i ness wi th most of the emergency water programmes substantially reduced the value of the benefits during the drought. But even where the borehol es and other improvements in supply were 1 ate, they have probably contributed to the long-term benefit of the population. It is likely, though, that such longer-term benefits could have been achieved more cost­effectively had the activities been undertaken in a carefully planned way as part of a normal development programme (for example in Lesotho) rather than as an emergency intervention. The combination of food-for-work, as a relief measure, with construction of dams to improve village-level water supply (Christian Care, Zimbabwe) is also an example of successfully combining immediate relief with longer-term goals that will command considerable local support in a crisis.

6.17 The health sector was not identified as a priority area for ODA support in its initial assessment but, given the outbreaks of drought-related diseases, perhaps preventative measures should have been given a higher priority. The lack of integration between emergency water and health measures is also a less satisfactory aspect of the drought response.

6.18 Appropriateness of the assistance provided, including the balance between financial aid, food aid, funding of NGOs and technical co-operation input: The innovative use of financial aid as BoP support was a very positive action. Such a response is especially appropriate where the recipient government has the capacity and commitment to organise an effective national disaster response, but

57

Conclusions Southern African Drought

where financing the response would involve economically damaging foreign-exchange costs. Assisting in this way, ODA also probably contributed to galvanising international support, for Zambia and Zimbabwe in particular, to 1 imit the negative impacts of the drought on their medium-term economic strategies.

6.19 Food aid was correctly targeted on Mozambique where the drought exacerbated the war's effects. However, within Mozambique the bulk of ODA food aid was distributed in areas other than those most severely affected by the drought. This reflected ODA's establ ished practice of inviting 'bids' from famil iar trusted implementing NGO partners even though these efforts were not concentrated in the most severely drought-affected areas. As appropriate information is lacking it is not possible to determine whether these allocations were part of a coherent overall response or whether the pattern of distribution diminished the potential overall international effort in the most seriously affected areas.

6.20 The concentration on the proviSion of cereals aid may have resulted in ODA's implementing partners distributing rations that were nutritionally unbalanced and that under certain circumstances could result in increased risk of micronutrient deficiency diseases in all age groups and PEM in weaning-age children. The information required, however, to determine whether this was so in the areas where ODA food aid was distributed was not available.

6.21 A strength of ODA's emergency funding is its considerable flexibility and simple requirements for disbursement. This was generally noted as resulting in rapid utilisation. Many NGOs also welcomed the freedom to begin operations on the basis of a commitment in principle prior to the issue of a letter of agreement. A further valued aspect of flexibility was in changing the actual utilisation of funds in accordance with changing needs in an emergency. Such flexibility, however, implies a need for further reporting as part of the responsibility for the use of public funds.

6.22 Co-ordination mechanisms available, their effectiveness, and how these affected ODA's actions: The international and regional early warning systems functioned well in providing relatively early information on the drought. The variable performance, however, of the governments of affected countries, and also of donors, in responding to warnings indicated that the co-ordination of actions in relation to such information is still less than satisfactory.

6.23 The improvised co-ordination by food aid donors was successful in initiating an adequate international response and support for the regional logistics operation, which ODA funded in a small way but also strongly supported in policy terms. It was an outstanding example of international co-operation.

6.24 The DESA process, which ODA also supported in a small way, was a less successful but still positive aspect of the international response. A shortcoming was the delay resulting from a perhaps over-ambitious attempt at regional co-ordination. The DESA was also unable to find a satisfactory way of integrating bilateral responses, direct to governments and through NGOs, with an appeal organised around food aid and the non-food proposals of UN specialised agencies.

6.25 Because of the reactive nature of ODA's funding in relation to NGO proposals, it was sometimes difficult for the officials in post to ensure that there was no duplication. The 'broker role' of some UK-based NGOs in seeking funding for their local partners from a number of donor sources, and their apparent lack of co-ordination, meant that it was extremely difficult for ODA to know whether it was funding activities with the highest priority. This reactive

58

Southern African Drought Conclusions

stance and the fact that most of the UK-based NGOs operating in Mozambique kept to their existing locations, prevented ODA from targeting its assistance on the worst affected areas to the extent that it would have preferred. Another issue for consideration is whether ODA should, in effect, through its financial grants, be meeting some of the ITSH costs of the food aid reI ief operations of other donors.

6.26 The lack of co-ordination between emergency water supply activities and related health measures was a particular weakness of programmes in both Lesotho and Zimbabwe.

6.27 More generally, NGOs retrospectively emphasised the need for stronger co­ordination amongst themselves in organising drought relief.

6.28 The effectiveness of monitoring and accountabil ity arrangements: The country case studies consistently highlight weaknesses in monitoring and reporting on the part of a high proportion of the agencies operationally involved in the drought response. Reporting was generally delayed, and in some cases perfunctory. This problem included both directly operational agencies and those providing funding for local partner agencies. In the case of the latter, the limited capacity of partner agencies is not a satisfactory explanation for inadequate reporting. The onus of responsibility rests with the UK-based grant­receiving agencies to ensure adequacy, if necessary through the provision of technical assistance to their partners. The lack of follow-up, or response, by ODA on reports was also widely cited by NGOs as a disincentive to do more than report in a minimal way. Where, however, the Post had an expl icit role in monitoring grants, eg WaterAid and LSM in Zimbabwe, a high standard of reporting was noted.

6.29 Weakness in monitoring and reporting, as well as in proposal s as al ready noted, is considered to be part of a wider issue of transparency in the use of and accountability for public funds. As the scale of the resources committed to relief operations has grown, this has been associated with the emergence of a quasi-autonomous system of relief organisations in which accountability is diffuse and diluted. International NGOs and consortia of NGOs, as well as international agencies, are drawing upon funding from many governments and the general public across the Western world. The scope for a single donor agency such as ODA to establish priorities is therefore severely constrained.

6.30 Subsequent to the period covered by this evaluation, ODA issued revised guidelines for NGOs applying for and in receipt of ODA funding for humanitarian assistance activities (ODA, 1994). These represent a welcome step in recognising monitoring and accountability concerns of the sort revealed by this study and, in general terms, how these might be addressed. In the light of the findings of this evaluation there is still scope for the further development of these Guidelines. For instance, it would be helpful to develop activity-specific checkl ists of questions to be covered by agencies when preparing project proposals and monitoring reports so as to reflect the very wide range of activities frequently undertaken in providing humanitarian assistance (eg general ration distributions, supplementary feeding programmes, water programmes, health programmes, agricultural rehabilitation and logistics activities). Similarly the provision of explicit unit-cost information related to the different types of activity is also needed, to show the cost-effectiveness of proposed and completed actions.

6.31 Evaluation success ratings: The terms of reference do not ask for an Evaluation Success Ratings Table (ESRT) for ODA's drought response. It was,

59

Conclusions Southern African DrollQiJt

however, subsequently agreed to consider the criteria for summarising the performanc,e of this regional emergency programme.

6.32 The ESRT method has been developed and refined in relation to development projects. But the OOA' s response to the Southern Afri can drought was qui te unlike a conventional development project in a number of respects:

a. This was a regional programme of activities involving a wide range of components in nine countries affected in different ways and degrees by the drought during 1991/2. In contrast, the ESRT covers particular aspects of performance of quintessentially the individual project.

b. OOA's response combined humanitarian with developmental objectives. However, in the absence of any explicit statement, it is difficult to assign relative importance to these concerns. For example, 40% of expenditure was provided as direct financial assistance to the Governments of Zambia and Zimbabwe's own emergency efforts, but also to help protect their longer-term development priorities;

c. OOA's bilateral response constituted a relatively small component, approximately 5%, of the overall international response. Consequently, apart from particular sectors or geographical areas within the affected countries where OOA's assistance may have been concentrated, it represented a minor part of international assistance, and in most cases it was difficult to determine separately the effects of OOA's contribution. The relevant assessment questions concern the overall response and, regarding OOA's contribution, whether it was complementary to and enhanced the overa 11 effort.

d. Emergency grants typically provided partial support to the rel ief programmes of UK-based NGOs and international organisations that were in turn often directly supporting only components of the activities of governments and NGOs in affected countries. For impacts, therefore, the chain of causation was often tenuous.

e. Emergency operations are by their nature transitory and, with certain exceptions such as emergency water interventions, often 1 eave no 'footpri nt ' .

f. Therefore, because performance monitoring and impact assessment were not accorded high priority by OOA in making grants or by its grantees in most of their actions, little can be inferred with confidence about impacts. The few exceptional cases of fuller impact assessment are not necessarily representative of the overall programme.

g. Ex-ante appraisal is a standard feature of the development project cycle, but in the absence of systematic appraisal for emergency operations, the benchmarks are typically lacking for ex-post evaluative rating.

6.33 To conclude, the appropriate performance criteria and the balance of performance criteria are different for emergency operations and development projects. Because of lack of performance monitoring and evidence on impacts, the qualitative assessment of OOA's drought response within the conventional ESRT is

60

Southern African Drought Conclusions

difficult and potentially misleading in two respects. First, it would provide a false sense of precision in the assessment. Second, it would focus only on some of the relevant criteria for assessing the performance of an emergency operation. In order to illustrate these issues, we have provided a commentary rather than a scale rating for the drought response in terms of the standard set of project performance criteria. This assessment is then complemented by a similar review of other criteria not usually included in the ESRT which are considered as potentially relevant to evaluating the performance of an emergency action.

6.34 Development project performance criteria

Cost over-runlunder-run: not relevant. Most activities involved fixed sum grants for specific purchases or the financing of activity within a fixed budgetary limit. Precise performance targets were not usually set. The built-in flexibility within emergency grants also allowed discretion in altering the balance of activities or specification of purchases.

Time over-runlunder-run: relevant only in certain specific cases, such as emergency water projects. A more appropriate criterion is timeliness of action (see below). In addition, the changing nature of needs is typically accepted as allowing a departure from the original timetable in an emergency request established on the basis of highly imperfect information.

Technical success: variable performance was indicated by the evaluation, but weak performance monitoring makes overall assessment difficult. In view of the uncertainties, however, high and uniform performance is not to be expected.

Cost-effectiveness: variable performance, but strongly shaped by the need for timely action.

Adherence to project conditions: variable performance in terms of financial reporting on grants. More generally, grants are made in a highly flexible way.

Promoting GDA's development priorities: emergency objectives are generally expressed in terms of preventing or limiting negative impacts.

Economic liberalisation: financial aid given successfully with regard to this objective. Otherwise this is not a direct priority of emergency programmes.

Enhancing productive capacity: only the overall drought response could be assessed on a regional, country and sometimes a more restrictive basis, as protecting the erosion of productive capacity at various levels.

Good governance: an i ndi rect obj ecti ve of ODA' s response was protecti ng the reform process in Southern Africa. Overall, the international effort was highly successful in this regard.

Poverty impact: an important objective of humanitarian assistance is to prevent impoverishment. The overall regional response was apparently successful, but, with few exceptions, data were lacking to assess the relative success in the most vulnerable regions and in the groups targeted by ODA grantees.

Human resources: education: an objective of the school feeding activities supported in Zimbabwe was preventing disruption of schooling. Only impressionistic evidence available for ODA-assisted programmes.

61

Conclusions Southern African Drought

Human resources: health: prevention of mortality, damage to long-run health, and disease and nutritional stress. Not initially a high ODA priority, but apparently a relatively successful aspect of the overall drought response to which ODA contributed. Lack of monitoring information from rel ief and health interventions assisted by ODA.

Human resources: children by choice: not relevant.

Environmental impact: in practice this is only an indirect objective of emergency responses to assist affected populations to cope without irreversibly damaging their environment. No direct evidence for ODA-assisted activities. However, other evaluations eg Tiffin and Mulele (1993) for severely affected Zambia, indicate no significant, irreversibly negative environmental impacts.

Impact on women: some emergency grant-ass i sted programmes were exp 1 i citl y targeted on women. In practice, however, this constitutes an indirect objective as a high proportion of vulnerable populations in several seriously affected countries were female headed households. There was an almost complete lack of monitoring information on gender-specific impacts in the three case-study countries.

Social impact: an important objective of emergency aid is minimising social dislocation. In these terms, the overall response was apparently relatively successful, but there was a general lack of micro-information for ODA-assisted grants.

Financial rate oj return; Economic rate oj return; Financial sustainability; Institutional sustainability; Overall sustainability: these are not relevant criteria for emergency actions. However, as some actions involved investment in water supply, the return to and sustainability of such investment ought to have been appraised more systemati ca 11 y. Emergency aid i nvo 1 ves the management of resource flows by a co-ordination of government and non-governmental organisations and multilateral agencies. Substantial resource flows invariably confer some benefi ts to the 'channell i ng' organi sati ons, resul ti ng in thei r strengthening relative to other organisations. For instance, a decision to channel resources through UK NGOs operating in parallel to agencies of the host government may strengthen their power in relation to the latter. Judgements as to whether this constitutes a positive or negative effect are invariably subjective.

6.35 Additional project performance criteria relevant to an emergency operation: Amongst the criteria missing from the standard list are the following:

Quality oj assessment of disaster impacts and related quality oj proposals for emergency action in terms of appropriateness and the other criteria listed below. In conditions of uncertainty many initial assessments and proposals will be retrospectively invalidated and highly variable performance is to be expected. The evaluators judged that the assessment of disaster impacts was satisfactory but the quality of proposals was extremely variable.

Timeliness: a quantitative assessment is required of impacts, as well as the management performance of grantees in preparing proposal s, ODA in processing proposals and grantees in implementation. As emergency requirements are criti cal, timel i ness overri des many other standard fi nanci al and techni cal considerations. 'An overall assessment of timeliness of a range of activities is impractical, but areas of variable performance included formulating and

62

S()uthern African Drought Conclusions

submitting requests. In terms of sectors, food distributions were relatively timely, seeds and tool distributions were variable and emergency water supply measures had little impact in relation to drought-created needs. ODA's own performance was satisfactory in terms of the time taken to review requests, and good in the provision of BOP support. Its performance in assessment, however, was variable.

Sens i t i vi ty to changing injormat ion and condi tions was a pos iti ve aspect of performance.

6.36 In order to provide summary success ratings for emergency operations the criteria for assessment would need to be restated in terms of a relevance for j udgi ng the effi cacy and effi ci ency of emergency acti ons. Accordi ng hi gher priority to monitoring and impact assessment is also a necessary pre-condition for more meaningful assessment.

63

SOUTHERN AFRICAN DROUGHT: EVALUATION

1. To review ODA's preparations, organisation and approach to the Southern African drought of 1992.

2. To consider the speed of response, both in overall terms and for individual requests.

3. To estimate the actual total cost of each of the major elements and to assess the value for money.

4. To describe the impact of each of the major elements.

5. To judge whether each of the major elements has or has not del ivered significant benefits in relation to the costs.

6. To review the appropriateness of ODA assistance provided, including the balance between financial aid, food aid and funding of NGOs and technical cooperation inputs.

7. To consider the coordination mechanisms available, how effective they were and how they affected ODA actions.

Background

It is suggested the consultant should spend 3-4 weeks on this exercise, with some assistance from an ODA (CSAD) officer. Most of the work can be done by studying files in ODA but a visit to the region should be included, possibly to cover Mozambi que (as a self-contained programme where the drought compounded the existing emergency); Zimbabwe (as a major recipient of ODA drought funding -Harare was also a regional centre); and Lesotho (as an example of a smaller recipient with perhaps less obvious needs).

65