Review of UNHCR Implementing Arrangements and Implementing Partner Selection Procedures

EVAL/08/97

SCOPE OF THE REVIEW

This comprehensive review of implementing arrangements, the first of its kind, is divided into two parts. The first part looks at the history of UNHCR's involvement with implementing partners, and examines the models that have emerged, comparing them and listing their advantages and disadvantages. It also reviews the current issues regarding implementing arrangements as raised by auditors, donors, Project Delphi and our partners themselves.

The second part of the review focuses on implementing partner selection. It analyses the extent to which UNHCR has a choice, looks at how choices are made in practice and suggests means for enhancing selection procedures. It also examines related topics such as evaluation of implementing partners, an NGO database, and capacity-building of local agencies.

Sources

Interviews were held with some 80 UNHCR and NGO staff (see Annex A). In addition, material was drawn from interviews conducted on the topic in 1993 by Lowell Martin, then Chief of UNHCR's Central Evaluation Section. The review also draws on considerable written material, including two previous examinations related to this topic, notes of workshops, a broad range of mission reports and country programme evaluations, Project Delphi and other policy documents, PARinAC conclusions and academic articles. Valuable support for the review was provided by the Division of Operational Support (DOS), at whose suggestion the review was undertaken.

Definitions

For purposes of clarity the following terms, used throughout the review, are defined:

• Implementing arrangements: The means by which UNHCR chooses to implement its programmes to provide assistance and protection to refugees and other beneficiaries.

• Implementing partner: Governmental, inter-governmental or non-governmental entities with which UNHCR enters into a sub-agreement to implement a project for UNHCR beneficiaries and which in principle bring additional resources of their own to meet needs which would otherwise have to be met by UNHCR.

• Operational partner: Agencies whose work is complementary to that of UNHCR and which may be assisting the same beneficiaries or working towards the same goals, but with which no sub-agreement has been signed.

The review was carried out by Fabrizio Hochschild, temporarily assigned to the Inspection and Evaluation Service as an Evaluation Officer, in co-operation with Lowell Martin and Robert White of IES. During the review, key documents were studied and in-depth discussions were held with some eighty UNHCR staff. Discussions were also held with personnel from a number of NGOs.

To accommodate the needs of various readers, two versions of the report are available: this comprehensive version, and a shorter "Synthesis Report".

OVERVIEW

1. The quality of service UNHCR provides to refugees depends primarily upon the effectiveness of the implementing arrangements adopted. But despite its critical importance, little thought has been dedicated to this subject.

2. No procedure exists to ensure thorough review of different implementing options at the outset of a programme or their evaluation in ongoing programmes. While guidelines exist for implementing partner selection, they are seldom applied. This means that crucial decisions affecting the lives of hundreds of thousands of beneficiaries and involving sums of tens of millions of dollars can be taken without research, without weighing options and on the basis of personal preference rather than more objective criteria.

3. The implementing arrangements used today have evolved largely through trial and error. The following basic models and sub-models can be distinguished. Their peculiar characteristics, advantages and disadvantages are analysed in the first half of this report.

4. The current trend is to work less with Governments and more with international NGOs and through direct implementation. Although direct implementation has increased, the number, currently over 640, and diversity of partners has also increased. The number of local partners has grown although the proportion of budget disbursed through them has not changed significantly. Other emerging tendencies in implementing arrangements are a renewed vogue - mainly in Eastern and Central Europe but with lessons that could be applied elsewhere - of assisting in the creation of agencies and the growing relevance of partnership with human rights organisations.

5. Major recent innovations in implementing arrangements are service packages, the use of which has remained exceptional, and extensive standby and secondment arrangements. The latter have greatly enhanced UNHCR's operational capacity and are credited with bringing a new spirit of pragmatism to programme implementation.

6. As well as credit for innovation, there has been criticism for lack of control over partners and weak financial management. Implementing arrangements have been the subject of increased scrutiny by auditors and donors and in many situations they have been found wanting. To correct this, there is consensus on the need for a more structured approach, in particular to implementing partner selection. However, in discussing possible improvements, the focus has thus far tended to be on financial pre-qualification requirements and audit issues. A broader approach is proposed in this report.

7. The measures put forward by this review to structure UNHCR's approach to implementing arrangements include the following:

• the introduction of formal requirements to justify the implementing arrangements adopted;

• the introduction of a selection procedure, which includes the application of mandatory prerequisites for working with UNHCR, pre-assessment of potential partners according to pre-defined criteria, and the involvement of a larger number of posts in the selection process;

• regular review and evaluation of ongoing implementing arrangements on the basis of agreed performance criteria;

• the establishment of a comprehensive implementing partner database;

• a more active approach to seeking agencies in sectors and locations where they are lacking;

• concerted and targeted capacity-building efforts in co-operation with international NGOs and other UN agencies; and

• the revision of the sub-agreement format to enhance its value as a contractual document and to allow for greater compliance with its sub-clauses.

8. Of the suggested measures, the most pressing is the introduction of an implementing partner selection procedure. The absence of a selection procedure makes UNHCR vulnerable to criticism, puts UNHCR in a weak position to resist the imposition of implementing partners, and, as indicated above, leaves scope for choices to be made without forethought and on the basis of inappropriate criteria.

9. The main argument put forward in defence of the current loose guidelines for implementing partner selection, is that there is no need for a more elaborate procedure as in practice, there is usually little choice. While there are often constraints, situations where UNHCR has absolutely no choice in implementing arrangements are rare. In most situations options do exist and choices are made. In certain situations there may be no choice, precisely because no effort was made to seek out or create choices. The possibility of implementation through commercial contract is, for example, at times simply not considered as it is thought to be too complicated.

10. The main purpose of a selection procedure is to ensure that all options are weighed and that the implementing arrangements adopted achieve the best possible relationship of value for money. The procedure proposed therefore emphasises cost effectiveness as a selection criterion and foresees total cost and total benefit of different implementing options being compared.11. Cost benefit analysis of implementing options is not consistently undertaken. Where a budget is available, cost effectiveness is not always a dominant selection criterion. Efforts to gain the maximum benefit for refugees while maintaining or reducing total costs are not usually sufficiently recognised.

12. Another weakness in current implementing partner selection is the fact that as indicated above, despite a long-standing policy of giving preference to local partners in selection, the overall proportion of local partners has not increased significantly. While in some circumstances genuine constraints exist to working with local partners, the decision not to work with them is at times the result of convenience or prejudice.

13. Working with local partners (contractors, NGOs, and government departments) has been described as enlightened self-interest; they are almost always less costly than international agencies, they are often closer to the beneficiaries, they remain and they allow UNHCR to make a more permanent contribution.

14. A number of measures are proposed to encourage more implementation through local structures. The selection procedure described foresees assessment of implementing options beginning with local agencies and makes contribution to local structures a formal selection criterion. The report also proposes more concerted efforts at capacity-building by encouraging mentor arrangements between international and local NGOs.

15. UNHCR's approach to implementing partners has not kept pace with changes in the context in which UNHCR works. The size of programmes has increased significantly and there are many new actors. Traditional mandates and responsibilities are being questioned. Roles have grown more fluid and are in flux. Donors have experimented in becoming operational. The role of UN organisations and Governments in the implementation of humanitarian and development programmes has declined. The role of NGOs in implementation and their importance as a lobbying force have greatly expanded. In the allocation of aid funds greater emphasis is being placed on cost effectiveness and the ability to co-operate with others, in particular NGOs.

16. Attitudes towards co-operation are not always constructive. There are some outstanding examples (such as Bosnia, Armenia, and Ngara) of programmes where UNHCR has most effectively combined the efforts of a broad variety of actors. Coexisting with these successes is an occasional - and damaging - inclination to institutional arrogance. In taking the credit for co-ordinating their efforts, UNHCR should recognize that its degree of dependence on its partners can sometimes be underestimated. One aspect of this is a tendency not to sufficiently recognise - despite PARinAC - the importance of NGOs in influencing the political and funding priorities of Governments. (An indication of this power is the ability of NGOs to block UNHCR attempts at private sector fund raising, while at the same time acquiring a growing portion of public funding.)

17. Co-operation requires mutual understanding. This can be furthered through measures such as staff exchanges, secondments, NGO internships and co-location in the field. While there are isolated examples of such initiatives, they could be more systematically encouraged.

18. While there is a necessity to co-operate more and better, sight should not be lost of the ultimate purpose of the partnerships which result: maximising the resources available to protect and assist UNHCR's beneficiaries. This report, it is hoped, will play a role in contributing to that goal.

SUMMARY OF RECOMMENDATIONS

19. A number of recommendations appear throughout the review, and are presented here in summary form:

Part I: Review of Implementing Arrangements

• Official counterparts should always be required. Ideally UNHCR should work through distinct structures for policy setting and programme co-ordination on the one hand and implementation on the other. The establishment of large, special structures should be avoided.

• Government structures to implement refugee programmes should be ideally staffed with seconded personnel who retain the right to return to their original departments. If feasible, their base salaries should be paid by the Government with UNHCR only paying incentives and travel allowances. If not feasible, UNHCR should pay salaries at government rates and ensure that contracts have explicit time limitations.

• UNHCR should only create agencies after other options have been exhausted. Then efforts should be made to expand an existing entity rather than starting something new. Efforts should also be made to plan phase-out or alternative funding sources from the outset.

• Field staffing policies act as a constraint to direct implementation, and commercial contracting procedures can discourage use of commercial contracts. Their adequacy should be examined.

• Direct implementation by UNHCR should remain restricted to the circumstances indicated in the Manual with a further criterion being added: direct implementation should be undertaken where this will result in considerable cost saving to UNHCR.

• The formula to calculate UNHCR's contribution to NGO Headquarters' overhead costs should be simplified.

Part II: Improving Selection

• Where UNHCR has no choice of partners and the implementing arrangements imposed are abusive, the continuation of the programme should be questioned.

• Headquarters could play a more active role in seeking NGOs where they are lacking.

• A more structured approach to implementing partner selection is required with the following elements:

• Close scrutiny of implementing arrangements should form an integral part of all operations planning. First assessment missions should include a thorough examination of implementing options within their Terms of Reference. The Country Operations Plan should include a section in which the field is required to justify the implementing arrangements adopted.

• As part of the selection process, potential partners should be pre-assessed according to pre-defined criteria adapted to local conditions and the choices at hand. These could be applied through a questionnaire. The level of scrutiny partners are subjected to prior to selection should vary according to the size of the project.

• Selection criteria should include minimum legal and financial prerequisites for working with UNHCR. In addition to those in the UNHCR Manual, criteria would emphasize technical capacity, suitability from a protection point of view and contribution of the agency to local structures. The contribution of an NGO, as a selection criterion, should not be treated in isolation, but should be taken into account in calculating the overall cost effectiveness of an implementing option.

• More emphasis should be placed on overall cost effectiveness as a selection criterion. All implementing options should be subjected to cost benefit analysis.

• The weight of different criteria in the selection process should vary according to the stage of the operation. The criterion of speed, for example, should be emphasized in emergencies and working with local structures in reintegration operations.

• Where many NGOs are present, where the contribution of agencies is negligible, where needs are well defined and where UNHCR has difficulty applying its own criteria due to pressure from donors, projects should be put out to tender and bids sought from agencies.

• Finance Officers and technical specialists, including Logistics Officers, should be involved in pre-assessment and selection of partners. Representatives should be encouraged to assume greater responsibility for implementing arrangements.

• Implementing partner performance in UNHCR-financed projects should be regularly evaluated on the basis of pre-agreed criteria and performance indicators. In the evaluation process, partners should be given the opportunity to highlight any constraints encountered in implementation and to comment on the quality of UNHCR support and guidance. More elaborate evaluation should be undertaken for larger projects.

• Implementing arrangements should be the subject of regular review. The Terms of Reference for inspection missions and country specific evaluations should include examining the appropriateness and cost-effectiveness of implementing arrangements.

• Sub-agreements should be reviewed as part of Delphi changes to programming instruments. Formats should vary according to the size of a project. A simplified format should exist for small projects. The review should also aim at making the budget structure more compatible with the budget structures used by partners. Reporting requirements should be re-examined. Further suggestions should be sought from partners.

• Information collection on implementing partners at Headquarters should be centralised and a comprehensive database established - if sufficient personnel can be identified to ensure it is kept up to date.

• To allow UNHCR to work more with local agencies, capacity-building should be undertaken systematically through a variety of means:

• Adopting a more proactive approach to meeting the equipment and training needs of local agencies.

• Providing more and better targeted training in programme and financial management. Providing technical training using staff available under standby arrangements.

• Encouraging mentor arrangements between international and local agencies.

• Undertaking comprehensive, longer-term capacity-building of targeted agencies in crisis- prone areas as joint projects with other UN agencies.

PART I: REVIEW OF IMPLEMENTING ARRANGEMENTS

THE EVOLUTION

The Statute, the 1950s and 1960s

20. UNHCR's Statute (General Assembly resolution 428 (V) of 14 December 1950) gives little guidance on who UNHCR should work with. Paragraph 1 of the Statute states that UNHCR should assist and protect refugees "by assisting governments and, subject to the approval of the governments concerned, private organizations." Paragraph 10 authorizes the High Commissioner to administer funds and specifies that he/she "shall distribute them among the private and as appropriate, public agencies which he deems best qualified to administer such assistance." In the absence of more precise guidance, implementing arrangements have largely evolved as a result of trial and error.

21. In the 1950s in Europe, Governments looked after most of the material aspects of the refugees' needs and UNHCR work focused mainly on protection. In implementing a limited number of projects, UNHCR functioned largely like an institutional donor. UNHCR's main partners were often church-linked (e.g., the World Council of Churches (WCC), the Catholic Welfare Conference (CWC), the International Catholic Migration Commission (ICMC), and the American Distribution Commission (ADC)) but not exclusively (e.g., the International Civil Service Commission (ICSC)). UNHCR also worked with smaller, national European NGOs.

22. When UNHCR started operating in Africa in the early 1960s, it could no longer depend on asylum Governments to provide relief, and NGOs were more scarce. Other UN agencies (such as the International Labour Organization in Zaire and Burundi) became important implementing partners. The League of Red Cross (today the International Federation of Red Cross and Red Crescent Societies (IFRC)) also became a major partner. In local settlement projects, UNHCR worked with a limited number of agencies who were considered to have development expertise, such as the Association Internationale de Développement Rural (AIDR) and the Lutheran World Federation (LWF).

23. It is estimated that in the mid-1960s UNHCR had between 10 and 20 implementing partners, of which half were large international NGOs. According to one staff member, at the time there was an unstated preference for large international NGOs, Governments being seen as too bureaucratic. Implementing agreements were signed mainly at Headquarters.

The 1970s and 1980s

24. The 1970s saw the beginning of large programmes and large offices. The diversity of beneficiaries grew as did the geographical range of operations. UNHCR began working on a large scale with returnees, with urban refugees as well as with internally displaced persons. This led to a greater diversity of implementing arrangements. The 1970s also saw the birth of a large number of non-governmental agencies in North America and Europe.

25. Many of the implementing arrangements UNHCR works through today have their precedents in the 1970s and have evolved little since then. In Central America, UNHCR worked through a large number of local NGOs, and NGOs with a vocal concern for human rights. In Thailand in 1979, faced with the Cambodian influx UNHCR became operational for a limited period on an, until then, unprecedented scale. In a programme managed by the wife of the US Ambassador, UNHCR used volunteers to help build Sakeo and Khao-i-Dang camps and in a forerunner to service packages, donors seconded engineers and other sector specialists. In the mid-1970s in Bangkok, UNHCR also started the practice of creating an agency to circumvent staffing ceilings.

26. In the 1970s and early 1980s, UNHCR increasingly found itself working in countries where refugees were considered politically sensitive and a potential threat to national security. Concerns about sovereignty influenced implementing arrangements. The role of government structures in implementation expanded and many large governmental agencies came into existence: Commissioner for Refugees (COR) in Sudan, National Refugee Commission (NRC) in Somalia, Comisión Mexicana de Ayuda a Refugiados (COMAR) in Mexico, Commissioner for Afghan Refugees (CAR) in Pakistan (1979/80), and the Administration for Refugee and Returnee Affairs (ARRA) in Ethiopia.

27. After UNHCR's expansion in the 1970s, in the early 1980s attempts were made to systematically reflect upon and adopt standard procedures to regulate dealings with implementing partners. From the early 1980s onwards an increased number of agreements, now called sub-agreements, were signed at the field level. At the urging of the auditors, standard sub-clauses were introduced, which were not revised until the early 1990s. In 1983 a study was undertaken entitled "Etude préliminaire sur l'évaluation quantitative des partenaires opérationnels du HCR dans la mise en oeuvre de projets pour les réfugiés". The results of the study are reflected in the relevant parts of this review.

28. In early 1984, the first guidelines were issued entitled "Co-operation with non-governmental organizations as implementing agencies for UNHCR-funded projects" (UNHCR/IOM 24/84 FOM/26/84) with the aim of introducing "a more consistent and planned approach to co-operation with non-governmental organizations". These were not superseded until the guidelines contained in Section 5.1 of Chapter 4 of the UNHCR Manual were drafted in 1991/92.

29. The 1984 guidelines raise the same issues under scrutiny today. If they had been followed in practice, it may have obviated the need for this review. They emphasize that implementing arrangements "should derive from the needs and objectives of the project and not vice versa". They introduce a basic selection procedure which includes clearance by Headquarters for all new agencies and foresees field offices "thoroughly assessing" potential agencies. They give preference to local organizations over international NGOs and foresee the possibility of strengthening their administrative and technical capacities inter alia through training and internships in UNHCR. They acknowledge donor preferences for their own NGOs but advise offices to "resist external pressure to appoint agencies that otherwise may be unsuitable". They foresee the NGO Unit collating information on agencies under a number of headings and undertaking, together with the relevant Bureau, joint assessments of agencies.

30. In 1986, the change of High Commissioner led to the initiation of an effort to evaluate all NGO partners. Evaluation missions were undertaken by the NGO Unit to Thailand and Pakistan. A single-page evaluation form was designed. This attempt reportedly foundered due to NGO resistance to the exercise and the limited support it met with in other parts of the Office.

31. The financial crisis in 1989 led to emphasis on the concept of partnership, which was subsequently to become a dominant theme in discussion of implementing arrangements. In its origins, partnership was primarily meant to be a means of saving costs by devolving tasks to other agencies and by increasing agency contributions to UNHCR programmes. The Executive Committee Working Group established during the crisis made a number of recommendations in this vein.

1990 - Today

32. The High Commissioner at the 1992 Executive Committee session reiterated the theme of partnership as a means of burden-sharing. In setting out her strategy of prevention, protection and emergency response, she emphasized that the "magnitude of the challenges clearly exceeds the capacity of HCR alone" and called for a "vigorous renewal of partnerships." This theme was echoed again a year later when she called for a "pentagonal partnership strategy" at the 1993 Executive Committee session.

33. The series of large-scale emergencies in the early 1990s and the search for new partners prompted the first significant innovations in implementing arrangements since the 1970s. Standby arrangements with NGOs and the use of military forces to support humanitarian programmes were pioneered in the Gulf crisis. In Bosnia, UNHCR, making use of seconded personnel, undertook direct implementation on a scale that had not been attempted before. In Zaire, when traditional implementation mechanisms proved overstretched, UNHCR developed the concept of service packages.

34. The early and mid-1990s also witnessed UNHCR operations being initiated in Eastern and Central Europe. Many new partners were engaged. In many of the countries government structures were being built up and there was no tradition of NGOs. This led UNHCR to focus for the first time on capacity-building of partners as an objective in its own right.

35. As the statistics in Table 1 and 3 (Page 12) indicate, the pace at which UNHCR takes on new partners today has outstripped the pace of budget increases. This reflects not only the increasing number of NGOs on the relief scene, but also the broader range of activities and geographical locations in which UNHCR is involved. It also reflects the growing number of local partners.

36. The tendency to move away from working with government structures and to implement more directly and more through NGOs is apparent from Tables 2 and 3. There are various reasons for this: the growing role of NGOs (see paragraphs 40-44 below); the increased involvement of UNHCR in emergency situations where government structures have collapsed or are absent; and the need, due to increased possibilities for early return, to work with flexible structures that can expand and contract easily.

PARinAC

37. The PARinAC exercise, launched in 1993 and preceded by a more restrictive consultation process, defined the framework for partnership with NGOs. The emphasis shifted from partnership as a means of saving costs (see paragraph 31 above), to partnership as a strategic alliance to assist and protect refugees and resolve refugee crises. Underlying this was an awareness of the need to develop a more equitable and open relationship, with NGOs playing a stronger role in policy setting.

38. After six regional meetings in Caracas, Addis Ababa, Tunis, Bangkok, Budapest and Kathmandu, PARinAC culminated in a Conference in Oslo from 6-9 June 1994. The Conference made 134 recommendations reflecting in some detail common concerns and policy objectives in protection, assistance and programme management.

39. Generally PARinAC is better viewed outside UNHCR than within. Within UNHCR it is sometimes said that UNHCR carries the greater burden in implementing the recommendations, and that PARinAC failed to address crucial financial and administrative issues. By contrast, academic and NGO observers quote PARinAC as a welcome sign of a UN institution realizing the growing importance of the NGO community and attempting to join forces.

The Growing Role of NGOs

40. The outbreak of an unprecedented number of large-scale emergencies in the aftermath of the Cold War led to the quadrupling of relief budgets. Seen as more efficient and less controversial, the lion's share of increases in aid funding was spent through NGOs in preference to multilateral channels. According to OECD figures, NGOs have now overtaken multilateral institutions as the second largest source, after bilateral funding for development and relief assistance. Collectively they provide over eight billion dollars in aid.

41. As a result of these increases, a greater portion of NGO funding now comes from government sources. According to one estimate, the proportion of funding received by NGOs from Governments rose from an average of 1.5 per cent of income in 1970 to some 40 per cent today.

42. In a related development, donors have significantly increased earmarking for funding of their own NGOs, which is disbursed through multilateral organizations. In Bosnia-Herzegovina in some cases multi-million dollar projects have had to be tailored to meet the requirements of a specific NGO (see paragraph 215 in Part II) in order to allow UNHCR to access funds from a specific national or intergovernmental donor. Elsewhere, NGOs have played an important role in mediating with their Governments to help UNHCR get large funding allocations which they would not otherwise have obtained.

43. Increases in funding and the mushrooming of emergencies also led to the creation of a large variety of new local and international NGOs. In response to growing numbers and in some cases dubious performance of NGOs, a few larger agencies are spearheading moves aimed at their self-regulation. The IFRC , WCC and others have formulated a "Code of Conduct" setting out basic principles for work in disaster relief. Efforts are also underway to gain agreement on a set of common basic, technical standards and minimum provisions for beneficiaries.

44. Parallel to the creation of many new NGOs, the larger western NGOs have grown further. NGOs like Médecins sans Frontières (MSF), Oxfam, CARE, World Vision and Catholic Relief Services have grown into large organizations, each commanding annual budgets of hundreds of millions of dollars. They have become more specialized and highly professional, rivalling or superseding UN agencies in technical expertise and capacity. They attract the largest proportion of government funding of NGOs. It is estimated that some 20 European and North American NGOs receive approximately 75 per cent of all public funds spent in emergencies.

Emphasis on Cost Effectiveness

45. The high profile of humanitarian operations and increases in aid spending have led to greater public scrutiny of aid allocations. There have been calls for less bureaucracy and aid departments have been compelled to cut back. As the political criteria of the Cold War period have become obsolete, effectiveness and financial management capacity have grown more important as criteria in the allocation of funding. Increasingly donors also seek greater visibility for their funding.

46. Throughout the UN system the issue of accountability for the increased amount of funding spent through implementing partners - and thus not subject to the same stringent rules as direct UN expenditure - has gained prominence. Two studies by the UN Joint Inspection Unit (JIU) also focus on this issue.

47. As an indication of changing donor priorities, at ExCom calls for closer partnership are increasingly being supplemented by calls for more control and better financial management of partners. Emphasis on effective spending and efficient financial management has also led some donors to propose that UNHCR put projects out for tender and seek bids from NGOs.

48. Amidst its operational successes, weak financial management is proving to be a vulnerable point for UNHCR. In 1993 General Assembly Resolution 48/216 explicitly criticized UNHCR's financial management practices. Audit recommendations (see paragraphs 163-171) have highlighted the same deficiencies year after year in the management of implementing partners, culminating in a threat in 1996 to qualify the accounts.

Changing Roles

49. In the post Cold War period, traditional roles are being questioned and have grown more fluid. A number of government agencies that traditionally acted as institutional donors have experimented in becoming operational. The British Overseas Development Agency (ODA) and OFDA DART in the Gulf crisis, and the European Community Humanitarian Office (ECHO) and the French Ministry for Health and Humanitarian Affairs in Bosnia, are examples. Elsewhere NGOs have become involved in roles traditionally assumed by Governments or multi-lateral institutions such as peace-making; the Italian NGO, Comunità di Sant'Egido, who played the key role in the Mozambican peace process, is one example.

50. As some of the traditional advantages of working through UN agencies grow less relevant and others are increasingly being matched by NGOs, UN agencies are under pressure to redefine their specificity and comparative advantages. Co-ordination responsibilities are much sought after as they provide a raison d'être. Agencies are also under pressure to be more operational as a means of maintaining credibility. UNHCR in Bosnia gained credit with Governments and NGOs alike by taking on a role (among others) traditionally played by NGOs.

51. Roles are evolving and mandates are no longer considered sacrosanct. There are many varying views about what the future holds - from downsizing of the UN to creation of a super UN agency combining all humanitarian functions. Whatever roles emerge, cost effectiveness and financial management capacity as well as the ability to cooperate effectively with partners, in particular NGOs, are likely to grow more important as criteria in the allocation of aid resources.

TABLE 1: 1983-1995 Type of Partner by numbers

	1983	1991	1993	1994	1995
International NGO	30	72	127	124	128
Local NGOs	20	193	291	318	336
Government Agencies	40	113	145	145	154
UN Agency/IOM	9	8	10	8	11
Total No. of partners	89	386	573	595	629

TABLE 2: 1991 - 1995, Type of Partner by Level of Expenditure

(as a percentage of overall expenditure through partners)

	1983	1991	1993	1994	1995
International NGO	25%	12%	22%	41%	40%
Local NGOs	25%	31%	29%	24%	28%
Government Agencies	55%	49%	42%	30%	26%
UN Agency	5%	4%	4%	1%	3%
IOM	?	4%	3%	3%	3%

TABLE 3: Proportion of Overall Budget Spent through Implementing Partners

	Expenditure	Through Partners	%
1991	772 million	330,000	46
1992	1017 million	340,000	34
1993	1104 million	433,000	39

REVIEW OF MODELS

Working with Governments

52. The level of government involvement in implementation varies considerably from programme to programme. While in some countries, such as China or Iran, the Government implements the whole of UNHCR's assistance programme, in other countries, like Honduras or Egypt, the Government plays a minimal role.

53. While the norm is to work through a structure linked to the Ministry of Interior or Social Welfare or their equivalents, there is considerable variety in the type of government mechanisms used for implementation. In Dubai, the UNHCR programme was implemented through an expatriate economic advisor to the Head of State. In the Philippines, the UNHCR programme is partially implemented through the military. In Thailand, UNHCR had different government partners for different refugee groups.

54. From among the many models that have emerged, two basic approaches to government implementation can be distinguished:

• The creation of a separate, large government structure which implements all or most of UNHCR's programmes in a variety of sectors; and

• The use of existing government structures, such as line Ministries, together with a lean co-ordinating body which oversees implementation but is not operational itself.

The advantages and disadvantages of these two basic models are examined further below. The degree to which UNHCR has a choice or can influence the structures established is examined in Part II in paragraphs 202-226.

55. As indicated in paragraph 36, the statistics indicate a steady decline in the proportion of assistance UNHCR channels through Governments. The preference for working with Governments is premised on the notion of government capacity and control over its territories. Increasingly, however, UNHCR finds itself working in situations where this does not apply.

56. While in some situations the channelling of assistance through government structures may be impossible or inappropriate, there is always a need for an official counterpart to facilitate the programme and to ensure security. The primary responsibility for assistance and protection to refugees and displaced rests with the host country regardless of the nature of its authorities. To ensure the involvement of official counterparts, in Afghanistan UNHCR worked with provincial authorities, while in Liberia and Bosnia UNHCR worked with municipal authorities or the factions controlling a given area.

57. Other situations have arisen such as in Nepal or Honduras, where the Central Government is in full control of its territory, but has left UNHCR with a very broad scope of responsibility in dealing with refugee affairs. Even in these situations it was considered expedient to encourage the creation of a small government Office dedicated to refugee matters to ensure government involvement and facilitate programme implementation.

Large, special structures

58. Special structures to implement refugee projects were established in what traditionally were some of UNHCR's largest and most intractable refugee problems. COR in Sudan, the NRC in Somalia and the Office National d'Assistance aux Réfugiés et aux Sinistrés (ONARS) in Djibouti were set up for Eritrean and Ethiopian refugees; CAR in Pakistan and the Bureau of Aliens' and Foreign Immigrants' Affairs (BAFIA) in Iran handled Afghan refugees; ARRA in Ethiopia assisted Sudanese refugees, and COMAR in Mexico focused on Guatemalan refugees. In some cases, such as with COR, the establishment of a large special government agency reflected the Cold War expectation that the refugees would stay for a long time. Elsewhere, notably in Pakistan, special structures were created precisely to emphasize the temporary nature of the refugee presence.

59. Under certain conditions, the establishment of a single, separate structure dedicated wholly to the implementation of refugee projects can be the simplest and most efficient means of meeting the needs of refugees. This is the case where there is a large influx, where the Government has a restrictive attitude towards NGOs and where line Ministries lack the capacity to expand rapidly to meet emergency needs. Many believe that in Pakistan, for example, only through a body like CAR could such a big and comparatively effective programme have been mounted.

60. The creation of a large special agency may also help in fostering an open attitude towards refugees. A large special agency dealing exclusively with refugees may be more aware of UNHCR protection concerns than other government departments with a less direct interest. Some believe that COR helped maintain an open door policy in Sudan, for example. Another advantage of special structures is the convenience of having a single government interlocutor for all matters relating to refugees, which may facilitate policy consistency in programme implementation.

61. One problem caused by creating special government structures is the potential for duplication and overlap with other government departments. In Sudan, COR schools existed in parallel to Ministry of Education schools. To ensure adherence to national standards some dealings with line Ministries are always required. This was the case in Somalia where at one point UNHCR was funding special refugee units in line Ministries and the NRC to undertake overlapping tasks.

62. A related problem is the tendency for large, special structures to wish to monopolize UNHCR resources. Other government departments and NGOs can come to be viewed as competition. COR resisted efforts to hand over health programmes to the Ministry of Health and discouraged informal refugee schools, led by refugee elders. In Somalia, according to a staff member there at the time, it took starvation of refugees to allow UNHCR to switch responsibility for logistics to CARE.

63. A further disadvantage is the tendency for such structures to grow independently of needs. In Pakistan at one point UNHCR was funding 22,000 staff, and 70 per cent of the programme budget was spent on salaries. Such large structures are inevitably slow and inflexible as well as costly. In 1989 the per capita cost to UNHCR of looking after refugees in the Sudan was between 5 and 10 times more than in Uganda and Zambia.

64. Frequently large structures prove inadequate at reporting and due to their size, hard to hold to account. COR, despite a large team of UNHCR-financed accountants, did not report satisfactorily. COMAR in Mexico and BAFIA in Iran have also been deficient in this respect. Lack of accountability can increase possibilities of corruption. A 1987 evaluation of the Pakistan programme concluded that corruption afflicted refugee administrations worse than other areas of Government, because resources were more abundant and the final beneficiaries were not nationals.

65. A major disadvantage of specially created structures is their exclusive dependence on UNHCR funding. This can lead to extraneous factors, such as the need to maintain a large bureaucracy, influencing government policy on refugees. Repatriation may be discouraged as this would take away the raison d'être of the agency concerned. Phasing out becomes lengthy and expensive.

66. As the previous paragraphs indicate, the disadvantages entailed in creating special structures largely outweigh the advantages. While there is a tendency to blame the structures themselves for the problems they create, the drawbacks are often predictable. A 1989 consultancy on the Sudan concluded: "It is easy to blame COR for working towards different aims, but UNHCR also bears much of the responsibility for having created a large and uncontrollable bureaucracy that now understandably has a major interest in defending itself and securing resources for the Sudan."

67. In some situations UNHCR, for the sake of simplicity and coherence of policy, still promotes the creation of one government body responsible for all refugee matters. This is the case in the former Soviet Union, where UNHCR has encouraged the establishment of centralized government Ministries or departments to bring together all migration-related issues. In contrast to the examples discussed in previous paragraphs, these bodies do not usually undertake sectoral implementation and the majority of staff are not usually paid for by UNHCR.

68. Where large government agencies are not totally dependent on UNHCR or where their responsibilities are curtailed, they can prove effective counterparts. In Bangladesh, the Ministry of Relief which assumed responsibility for refugees was highly experienced in the provision of relief and is judged to be an effective partner. In Southern Sudan, Project Management for Refugee Affairs (PMRA), an offshoot of COR with more limited responsibilities, also proved to be a co-operative and effective partner.

Lean structures and line Ministries

69. The implementation arrangements in Malawi, often quoted as exemplary, were established in a conscious effort to avoid the problems encountered with large special structures. The essential features of the Malawi model were:

• a small co-ordinating/policy-setting body, which held executive powers and was located in the President's Office;

• implementation largely through existing line Ministries and through commercial arrangements, the Red Cross and a limited number of international NGOs; and

• the secondment of sector specialists from line Ministries to UNHCR to undertake sectoral co-ordination.

70. The advantages of this approach are many: the cost of the programme was comparatively low; refugee assistance was carried out in line with national standards; good relations between NGOs and Government predominated; the programme proved relatively easy to phase out when the moment for voluntary repatriation came; and the programme also benefited nationals and helped build up the capacity of line Ministries.

71. A disadvantage of working through line Ministries is the difficulty sometimes of distinguishing between assistance going to locals and assistance going to refugees. A further disadvantage is the potential for confusion and added workload in dealing with many government counterparts. A third disadvantage is that line Ministries are sometimes more used to longer term development work and may not be able to adapt easily to meeting emergency needs.

72. For these reasons, working through line Ministries tends to work best in countries where they are present throughout the country and where there are high levels of organization. A comparable approach to Malawi worked well in Costa Rica. An initial attempt to do something similar in Somalia (which pre-dated Malawi), however, was less successful.

73. In general it has proven advantageous - as was the case in Malawi - to have distinct bodies undertake co-ordination and policy setting on the one hand and programme implementation on the other. Where there is no clear distinction in government structure between policy setting and implementation, as in the case of large, special structures, conflicts of interest may arise and UNHCR is left without recourse when there are problems with implementation.

74. The absence of a single, strong policy focal point in Government who can act as a UNHCR advocate is a major disadvantage. An arrangement where there are a variety of different government departments for different aspects of refugee policy (which has been the case in Kenya, for example) or where the main counterparts have limited authority and access, can significantly hamper a programme. To avoid this and to ensure direct access to the most senior policy makers, in Zaire UNHCR assisted in the establishment of "Cellule de Crise" under a Special Advisor to the President.

75. Government co-ordinating bodies can be a good solution where they have decision-making power and are under strong leadership. Both in Armenia and Georgia, refugee co-ordinating committees proved highly effective partners and transformed themselves into important government departments largely due to the strength of the personalities that led them. In Malawi and Benin, the co-ordinating bodies derived the necessary authority from being located in the President's office.

Staffing and salary issues

76. From previous experience it has become clear that a key element in government implementation structures are staffing and salary policies. Where provisions made at the outset are inadequate and the need arises to scale down, these issues can overshadow all other issues and become a major factor influencing government policy vis-à-vis refugees.

77. UNHCR policy on staffing and salaries in government agencies is reflected in standard clauses 6.1 to 6.7. of sub-agreements. The clauses state that the Government shall second staff to undertake UNHCR projects; that the Government will meet the basic salaries of seconded staff, and that seconded staff will have the right to return to the agency/department from which they came. Where the needs of the programme demand that additional staff are recruited, recruitment is to be undertaken in consultation with UNHCR.

78. In many programmes, most notoriously in Pakistan and the Sudan, these clauses were not strictly adhered to, engendering many problems. In both Pakistan and the Sudan there were large discrepancies between the staff that existed on paper and those actually undertaking a job. In both programmes it has also proved hard to reduce staffing levels and to phase out. In Pakistan, according to an 1987 evaluation, there were instances of government posts being sold.

79. In the case of the Sudan, COR paid salaries at higher rates than other government departments. This policy is sometimes pursued to attract competent personnel to remote areas where refugees are often located and to discourage other forms of income generation where salary levels are very low. However, these advantages may also be achieved by other means.

80. Various other possibilities exist:

• UNHCR pays a lump sum to the Government to meet all administrative costs including salaries. This is the situation in Iran.

• UNHCR pays through the Government the salaries of low wage staff who are recruited directly by the project. More senior staff are seconded and their base salaries are paid by the Government. UNHCR pays additional allowances for field living or travel. This was the case in Nepal, Benin and Bangladesh as well as many other countries.

• The Government meets all government staff salary costs. UNHCR pays only for other project costs. This was undertaken in Croatia as well as in various other countries in Central and Eastern Europe.

81. While the last option is obviously the most desirable, it is only practical where government infrastructure is developed and relatively well resourced. The first option has the advantage of ease of management but leaves UNHCR with no control over administrative expenditure. The second option is the most appropriate in many situations where the Government has less means.

82. In summary, to facilitate phase-out it is most desirable wherever possible to work with seconded staff. Ideally the Government will pay the basic salary costs of such staff in accordance with the relevant sub-agreement clause. Where this is not possible, salaries should be paid at government rates to avoid discrepancies. The payment by UNHCR of additional field living and travel allowances can provide a useful means of encouraging officials to undertake field trips and provide an incentive.

Working with Non-Governmental Organizations

83. As the statistics in Table 2 indicate, the larger portion of UNHCR's programmes is now implemented by international and local NGOs. NGOs now account for 65 - 70 per cent of UNHCR expenditure through implementing partners as opposed to 40 - 50 per cent in the early 1980s.

84. One of UNHCR's principal strengths is its relationship with NGOs. The added value of working with NGOs is considerable. While no reliable figures could be found, the amounts spent by NGOs on refugees and other UNHCR beneficiaries with funds raised from other sources, amounts to hundreds of millions of dollars.

85. Over the past five years the following agencies were UNHCR's main international, non-governmental implementing partners:

International NGO	Approx. Average Annual Level of UNHCR Funding (US$)
International Rescue Committee (IRC)	22 million
CARE International Belgium	8 million
Oxfam	7.2 million
MSF Belgium	7 million
Lutheran World Federation	7 million
MSF France	6 million
Adventist Development and Relief Agency (ADRA)	5 million
CARE Canada	4.2 million
Norwegian Refugee Council	4.2 million
Save the Children Federation (SCF) UK	4 million
Danish Refugee Council	3.5 million
CONCERN Ireland	3 million

86. In addition to their own funding, NGOs bring invaluable technical expertise and capacity. In Goma, Oxfam's water systems were generally judged more effective than those of the US Army. In the US, Europe and elsewhere, NGOs have acted as powerful protection advocates where UNHCR had to be more cautious about speaking out. They provide an indispensable information channel for both needs assessment and protection concerns. They have a knowledge of local conditions without which UNHCR in many places would find it hard to function.

87. NGOs have also proved to be powerful critics. At times there have been ideological differences, for example with MSF over assistance policy to Rwandese refugees in 1995. Criticism of UNHCR inefficiency, mismanagement, indifference and waste is a more constant theme and has contributed to bringing about improvements.

88. One of the main features of NGOs is their diversity. NGO partners range from agencies comprising a single person, vehicle and office, to organizations which in terms of staffing and budget rival medium size UN agencies. In outlook, work methods and capacity there are also huge variations between NGOs and in some cases, in different programmes within the same NGO.

89. NGO partners also vary over time. Oxfam started specializing in supplementary feeding and is now known for water. AIDR was a major UNHCR partner in the 70s and 80s and is now little known. Action contre le Faim (ACF), little known in UNHCR six years ago, is becoming an important partner. The Bosnia emergency brought into being many new agencies, such as British Direct Aid, which afterwards were taken on as partners elsewhere.

90. The diversity and independence which characterize NGOs also mean that some have less desirable features. At worst, they can include organizations that are amateurish, willing to put publicity before anything else, prone to proselytizing, uncooperative, or thinly veiled channels for inappropriate information gathering, political or commercial gain. Such shortcomings are not limited to smaller, new NGOs.

Local NGOs

91. UNHCR has a long tradition of working with local agencies. The level of involvement varies between programmes according to the availability and capacity of local partners. In Western Europe, local agencies are taking over functions which traditionally have been considered core UNHCR functions. In Eastern Europe, it has proved more difficult to identify local partners and efforts have been made to stimulate their creation.

92. Since 1984, it has been UNHCR policy to use local implementing partners in preference to international agencies. The importance of working with local NGOs is periodically reiterated (see, for example, PARinAC recommendations 36, 66, 75 and 88). As the statistics in Table 1 and 2 indicate, the reality has lagged behind the policy. UNHCR has taken on more local partners, but the proportion of budget spent through them has remained broadly stable.

93. Across the globe, Red Cross and Red Crescent societies predominate as local partners. In 1995, UNHCR channelled over US$ 25 million through 36 national Red Cross and Red Crescent societies. Although classified as NGOs, in some countries Red Cross and Red Crescent societies have semi-official status and can be designated by the Government to work with UNHCR (in Malaysia for example). Political ties of a national society elsewhere, such as in Burundi, have made co-operation difficult.

94. The IFRC is an important implementing partner of UNHCR in its own right and UNHCR has co-operated with it in countries such as Guinea, Malawi and Sierra Leone, in order to build up the capacity of national societies. This has happened in an ad hoc fashion and the IFRC has sought greater co-operation from UNHCR in co-ordinating allocation of funding to national societies.

Local partners versus international NGOs

95. It is frequently suggested that local agencies work best where they are associated with an international organization, as in the case of local Red Cross and Red Crescent societies and some church organizations such as CARITAS. The view also prevails that the best examples of collaboration with local agencies are in counselling and social service projects.

96. There is a general tendency in particular in emergencies (see paragraphs 274-276 in Part II) to favour working with international agencies rather than "risk" working with local agencies. International agencies are familiar and implicitly trusted. However, in some emergencies - Bosnia in early 1993, Benin and Chechnya are examples - the scarcity of international agencies compelled UNHCR to work predominantly with local partners, with positive results.

97. The advantages associated with working with local agencies are the following:

• Knowledge of local situation: Agencies rooted in local communities have better access to information and to the local authorities than international agencies. In Pakistan when access to Afghanistan was limited, international agencies and NGOs alike relied heavily on local Afghan agencies to gain information on needs and as a channel for assistance.

• Sustainability and constituency building: Local agencies have local constituencies. Through working with local agencies, UNHCR is making a more durable contribution to the local community. Local agencies remain whereas international agencies often leave after international interest has declined.

• More cost-effective: With local agencies there are no overseas Headquarters' overhead charges. Salary costs of locally-recruited staff are likely to be far lower than those brought from overseas. Local agencies can often also obtain goods and services at more advantageous rates. In Bosnia, the cost of using local agencies for community service programmes has been estimated to be a quarter of the cost using international agencies.

• Commitment: In particular where local agencies are engaged to look after related beneficiaries, they will often take greater risks to meet their needs than international agencies. In early 1993, UNHCR in Bosnia used local agencies to smuggle food to besieged populations when all other means were barred.

• More disposed to receive guidance: UNHCR contributions to a local agency often make up the larger part of its budget, allowing UNHCR a greater role in setting its priorities.

98. The constraints associated with working with local agencies are:

• Hard to know: Particularly in countries where UNHCR has limited experience, local agencies may be unfamiliar. It is not always easy to determine whether they are legitimate humanitarian agencies. In Goma, an agency established an orphanage with borrowed children to access aid funds.

• Can be partial or subject to manipulation: Local agencies are more vulnerable than international agencies to political pressure and manipulation. They are rarely neutral and may be strongly associated with one side of an ongoing conflict. Support to one local group and not another may have a divisive effect. Many local NGO representatives with whom international agencies had worked in Rwanda prior to the genocide, turned up in camps in Zaire. Although in principle desirable as partners, it became apparent that many of them had strong links with the Interahamwe.

• Unfamiliar with HCR procedures: Unlike international agencies, local agencies are often unaware of HCR budgeting and reporting requirements and ill-equipped to meet them. They tend to be less efficient in reporting than international agencies. Training in UNHCR programming, budgeting and reporting procedures is time- and labour-intensive.

• Limited size and technical capacity: Local agencies tend to be small, while UNHCR often requires partners capable of undertaking large projects. They do not have the same resources, technical expertise and experience as international agencies. Their enthusiasm to take on tasks in some cases may not match their ability to meet them. Building up capacity of local agencies is possible but time-consuming. In emergencies, time is not available.

• Potential for dependence on UNHCR: Where UNHCR contributions make up the major part of their budgets over a number of years, local NGOs can grow dependent on UNHCR and the continued presence of refugees. They also lack resources to cope with UNHCR delays in payments of instalments or deal with sudden budget cuts.

• Non-existent or unwilling to work with refugees: In particular in the former Soviet Union, local NGOs are scarce. Where they exist, they are often unwilling to work with refugees.

99. Political involvement of NGOs is neither limited to local agencies nor negative in all cases. There have been instances of international NGOs actively assisting in furthering the interests of their own Governments. Political involvement in the case of local agencies may mean a high degree of commitment to the beneficiaries and to the programme. The work of local agencies in Honduras, Armenia, Northern Iraq and Bosnia, while far from neutral, has been judged essential and in many cases undoubtedly brought aid to victims more effectively than was possible for international agencies.

100. In other situations, encouraging local agencies can mean encouraging a military cause. In Goma, for example, while capacity existed among the refugees to take over large parts of the water programme, this step was not encouraged due to the obvious danger of putting an essential resource in the hands of Interahamwe leaders.

101. As the dividing line between victims and combatants is often unclear, it can be hard to distinguish between agencies furthering a military cause and agencies primarily dedicated to helping the victims, albeit on one side. A judgement call is necessary to determine when working with local agencies will further strife or where it will primarily assist victims who can not otherwise easily be reached.

102. Due to the disadvantages associated with them, there is a tendency at times to overlook the capacities of local agencies. Even where time would allow for this, rarely are resources dedicated to the task of undertaking an in-depth evaluation of the implementing potential of local agencies (Government and NGO). It is often presumed that they are weak. This presumption can become a self-fulfilling prophesy as it usually entails the majority of resources being dedicated to international agencies.

103. While there are many places where local NGOs are scarce, local organizations of some sort exist almost everywhere. These can range from women's groups to boy scouts (who undertook inter alia the task of body burial in Goma) to the local authorities. All make potentially valuable counterparts. In Northern Iraq, UNHCR implemented a shelter material distribution programme effectively through traditional community leadership structures. In Bosnia, distribution was largely undertaken by municipal authorities.

104. Various methods exist to gain familiarity with local agencies and evaluate their legitimacy. In the late 1980s in Pakistan, UNHCR gave limited, pilot project grants to Afghan NGOs and undertook a large training programme in UNHCR budgeting and reporting procedures. At some expense, after a year the office was in a better position to determine the more appropriate from the less appropriate partners.

105. Efforts to build up local agencies pay off. The time and money required to undertake this are minimal compared to the savings that can be gained in the longer term. One means of capacity-building is by encouraging pairing arrangements with international agencies. This is examined in more depth in paragraphs 312-316 in Part II. Capacity-building can include help in how to gain funding from other sources to reduce the potential for dependency of local agencies on UNHCR.

Government versus NGOs

106. UNHCR's Statute is sometimes interpreted in such a way as to give preference to government departments as implementing partners. The Statute is ambiguous in this regard. While it indicates that UNHCR should seek the approval of the Governments concerned to work with "private organizations", it leaves it to the High Commissioner to select between Government or NGO depending on which she "deems best qualified" to undertake the task.

107. In general, the majority of those interviewed favoured NGOs over Governments as implementing partners, quoting the following advantages:

• Flexibility: NGOs tend to be smaller and have less rigid structures than government departments. They can adapt more easily to changes in policy or circumstances. Many also have enough resources to be able to cope with UNHCR delays in disbursement or sudden budget cuts. They are less constrained by political considerations (NGOs initiated the cross-border operation into Eritrea when this was taboo for all others including UN agencies).

• Increased leverage and accountability: UNHCR is usually better placed with NGOs to insist on adherence to reporting and budgeting requirements.

• Increased funding: In many situations NGOs make a significant contribution to a programme. International NGOs can also prove easier to raise funds for than government departments (this has been the case in Bosnia).

• Presence and specialized emergency expertise: NGOs have the expertise and capacity to deal with large-scale emergencies that may overwhelm under-resourced and less experienced government departments.

• Commitment: In many refugee situations there may be ideological differences between the Government and the refugees (e.g., Honduras) or the refugees are seen as a low priority. The attitude of NGOs may be more positive and more committed.

• Closeness to beneficiaries: NGOs may be present where government structures are weak or non-existent. They are often more likely also to have the trust of refugees.

108. On the other hand, working through government implementing partners, in particular where these are line Ministries, has the advantage of ensuring that the refugee programme is co-ordinated with national relief or development efforts and is undertaken in accordance with national standards. This is particularly relevant in local settlement and reintegration programmes (see paragraphs 277-280 in Part II).

109. Some other advantages of implementing through Government structures are: they are permanent, thus facilitating sustainability; it allows UNHCR to make a more lasting contribution and can also help to build up national capacity which may enhance emergency response in the future; and in many instances, in particular where the UNHCR contribution to salary costs is low, working with Governments can be more cost-effective than international NGOs.

110. Lack of familiarity or a prior poor experience of working with Governments can lead a UNHCR office to overlook government departments, even where they are present and potentially capable partners. In the health sector there is often a tendency to resort automatically to international NGOs. In Tanzania an agreement was to be signed with Pharmaciens Sans Frontières to maintain a central pharmacy, when the Government drew attention to its own structures to undertake this task - an arrangement that worked out well.

111. The problem for UNHCR is often how to ensure appropriate government involvement without the result being an open-ended commitment to support large bureaucratic structures. In Malawi as well as elsewhere solutions have been found. Appropriate staffing policies (see paragraphs 76-82) and explicit time limits on contracts of project staff can help in this regard.

112. In most programmes, a mixture of government and NGO implementing partners is seen as the most desirable solution. A number of tasks, such as refugee registration or camp administration, are by their nature the responsibility of Government. For many other tasks, in particular in emergencies, the additional skills and commitment of NGOs are indispensable.

Standby Arrangements

113. From 1991 UNHCR began entering into standby agreements with a range of organizations. These were aimed primarily at enhancing UNHCR emergency response capacity through the provision of personnel, technical experts and logistics support. UNHCR now has standby arrangements with the Swedish Rescue Board, Norwegian and Danish Refugee Councils, UNV's Humanitarian Relief Unit, Red R (Australia), Rädda Barnen, the Centres for Disease Control and the Russian Government agency EMERCOM.

114. Standby arrangements are acknowledged to have many benefits: UNHCR can now swiftly deploy a body of qualified staff to emergencies; seconded staff have rejuvenated UNHCR, introducing a more hands-on and pragmatic approach; technical personnel available under the agreements have also allowed UNHCR to implement more directly when this is required or to more adequately select and oversee implementing partners.

115. Some of the disadvantages of standby arrangements are related to their success: UNHCR has been unable to accommodate other organizations who wished to enter into such agreements; it has also proved hard to find deployment opportunities for seconded staff when there are no emergencies. A further disadvantage is that these arrangements implicitly favour international over local resources. This also has cost implications. Use of local personnel and expertise (in telecommunications for example) can be considerably cheaper.

116. Some have suggested that UNHCR should enter into more standby arrangements with the aim of ensuring implementation capacity in key sectors, such as health and water. This has not been pursued for a number of reasons: some NGOs are unwilling to commit themselves in advance; standby arrangements would reduce flexibility in choosing partners; and in the case of major emergencies, the NGOs in question (such as MSF, SCF, and Oxfam) are likely to be present anyway.

UNHCR-created Agencies

117. In a number of situations where Branch Offices have been unable to identify a suitable local or international agency, they have resorted to creating an agency. Typically such agencies undertake functions for which it is hard to find partners or which in other situations UNHCR implements directly. Such agencies are also used to circumvent restrictions on staffing levels.

118. Tasks undertaken by UNHCR-created agencies include status determination, managing resettlement programmes, individual case counselling, and running social services projects and programmes for urban refugees. Examples of such agencies include the Foundation in Support of Refugee Assistance in Thailand (FISRAPT) in Thailand, and Agency Voluntary Service (AVS) in Hong Kong.

119. The practice of creating agencies is not unique to UNHCR. There are a number of successful examples of agencies created by others. After all international agencies had left Liberia in 1990, the US Office of Foreign Disaster Assistance (OFDA) created an NGO to undertake food distribution and provide medical services. The United Nations Office for Co-ordination of Humanitarian Affairs in Afghanistan (UNOCA) set up several Afghan agencies to undertake demining projects. The International Committee of the Red Cross (ICRC) in Somalia in 1992 set up women's committees initially to run soup kitchens and subsequently other tasks.

120. UNHCR experience with setting up agencies has been mixed. The advantages associated with them are:

• Cost: UNHCR-created agencies have been described as "a cheap way to do protection." The cost of employing staff through created agencies is much less than if UNHCR had to employ them directly.

• Knowledge of local conditions: Created agencies have many of the advantages of local agencies.

• Modelled to UNHCR requirements: In the case of self-created agencies, UNHCR can insist on the agency concerned using UNHCR budgeting and accounting systems. As created by and dependent on UNHCR, UNHCR in principle has greater control and can direct the agency as required.

121. The disadvantages associated with self-created agencies are:

• Status of the staff is unclear: Staff sometimes doing the same tasks as those (usually) undertaken by UNHCR staff do not enjoy the same rights or protection as HCR staff.

• Special status can cause confusion and resentment: UNHCR-created agencies have a special link to UNHCR. To other agencies, this can look like patronage or corruption and cause resentment. When in Hungary UNHCR set up an NGO, this was seen by the Government almost as an unfriendly act.

• Loss of control: With time, the agency can grow less willing to accept guidance from UNHCR. In Mexico lawyers hired through a UNHCR-created agency, Servicios de Representación Professional y Técnica (SERTEC), ended up attacking UNHCR in court.

• Sustainability: Self-created agencies are usually totally dependent on UNHCR funding and thus hard to phase out. Often they are specialized in implementing UNHCR projects and of little interest to other donors. Phase-out of UNHCR-created agencies proved a serious problem in Central America. Elsewhere agencies (such as the Greek Refugee Council) intended to be self-supporting are still wholly dependent on UNHCR funding.

122. While the disadvantages have discredited the practice in many places, in Central and Eastern Europe UNHCR has pursued a policy of stimulating the creation of agencies. The aim is less to bypass staffing restrictions than to encourage the creation of what are considered essential elements of civic society and to create alternatives to implementation through Governments in former socialist countries.

123. In Georgia, for example, an NGO was established to stimulate the creation of other NGOs through small grants. A similar project is being undertaken in Bosnia through the Independent Bureau for Humanitarian Issues (IBHI). In Slovakia new agencies have been created with UNHCR encouragement. In a much praised example in Armenia, UNHCR assisted in the establishment of a human rights centre together with the Council of Europe and the Government.

124. Self-created agencies tend to work well where a number of conditions prevail:

• Where UNHCR clearly has no alternative: I.e., no other NGO or government agency is able or willing to undertake the task, and self-implementation is not feasible (see example of Metkovic warehouse below).

• Where they are set up on the basis of an already existing organization rather than being started from nothing: It is important that the agency has some raison d'être other than UNHCR. The work of advocacy organizations has been successfully expanded inter alia in Belgium and Serbia to undertake individual case work.

• Where the time-frame of operation of the agency is clearly defined from the outset or where efforts are made to ensure funding from other sources after UNHCR withdraws: In Tajikistan, a UNHCR-created agency was trained in fund raising to gain access to other donors.

125. The main reason behind the creation of many agencies is UNHCR staffing policies which prevent direct implementation by UNHCR even where this is the best solution. An example is the Metkovic warehouse in Croatia, set up in late 1992 by UNHCR as no other agency was available to undertake the task. At the height of the war, Metkovic operated 24 hours a day and shipped over half a million tons of relief supplies to Sarajevo and Central Bosnia. Initially all staff at the warehouse were recruited as UNHCR staff. Upon Headquarters' insistence, however, efforts had to be made to set up an agency to employ them.

126. Other UN agencies, such as UNRWA, have established a different body of rules and procedures for local project staff. This allows UNRWA to employ tens of thousands of project staff. Something comparable would obviate the need to meet every situation where a task has to be done and no partner can be found with an improvised approach, which does not always afford optimal conditions for the staff concerned. A re-examination of UNHCR field staffing policies in this light may be appropriate.

UN Agencies and IOM

127. UN agencies: As indicated in Table 2 above, in terms of expenditure, UN agencies play relatively minor roles as UNHCR implementing partners (in contrast to IOM). UNHCR's main UN implementing partners (as opposed to operational partners - WFP is UNHCR's main operational partner) are UNDP, UNICEF and UNV. Other agencies which have acted as UNHCR implementing partners include ILO, FAO, WHO, UNESCO, HABITAT and UNOPS.

128. UNDP plays an important role in representing UNHCR in some 20 countries where UNHCR has no presence. UNV, through a standby agreement for emergencies as well as through its regular volunteer programme, has played an increasingly important role in providing staff both for emergencies and regular programmes. 129. Over the past five years, on average UNHCR disbursed the following amounts:

129. Over the past five years, on average UNHCR disbursed the following amounts:

Through US$ (million)

UNDP 4

UNICEF 2

UNV 2

130. Despite regular calls for greater co-operation, some constraints have stood in the way of using UN agencies as implementing partners:

• Implementation capacity: With the exception of WFP and to some extent UNICEF, most UN agencies specialize in technical expertise. They advise Governments and oversee implementation of programmes rather than implement themselves.

• Development orientation: Many UN agencies are not geared for rapid emergency response. They specialize in development and work with different time-frames. The World Bank, for example, requires between two and three years to design a project.

• Cost: UN agencies charge overhead costs of 13/14 per cent. They are usually more expensive than NGOs.

• Availability of NGOs: Many NGOs are available with comparable expertise and capacity to some UN agencies. ILO was an important partner for UNHCR in many parts of Africa, but many of the tasks it would have undertaken are carried out by NGOs.

131. There is a tendency to place unrealistic expectations on other UN agencies and then complain when they are not lived up to. This can stem from limited knowledge of how they operate. Attempts to get UN agencies to take over a programme when UNHCR wishes to leave (for example, recently in Tajikistan) inevitably meet with limited success.

132. UN agencies have, however, proved valuable as operational partners, in particular in those countries where they have a large presence and where support has been sought in their area of expertise and capacity. In Vietnam, for example, UNICEF gave crucial support to the UNHCR programme by drilling wells.

133. In 1993 in former Yugoslavia an innovative attempt was made to put WHO in a position to co-ordinate all health aspects of an emergency programme, with UNHCR initially financing WHO. Views are mixed on the success of this venture. On the one hand WHO arrived late and proved relatively expensive as an implementing partner. On the other hand, it undertook the task with authority and expertise which would have been hard for an NGO to rival.

134. In many areas, the international character, unique mandates and expertise of UN agencies cannot be matched by NGOs. An example is the UNHCR. The High Commissioner for Human Rights is potentially a highly valuable partner for UNHCR, in particular in countries of origin, and it is generally thought that UNHCR assistance in building up their operational capacity should be reinforced.

135. The usefulness of Memoranda of Understanding with UN agencies has proved to be proportional to the level of detail with which they spell out a division of labour and respective obligations. The Memorandum of Uunderstanding (MOU) with WFP and a country-level MOU with UNDP in Mozambique have been praised as positive examples. Where detailed MOUs cannot be drawn up, this may be an indication of genuine limits of co-operation.

136. IOM: A detailed review of previous co-operation in the field may be appropriate prior to drawing up an MOU. This was not done with IOM, with which experience has been mixed. IOM was judged a valuable partner in Vietnam, in managing the Orderly Departure Programme. It has also been much praised for its role in evacuating third country nationals fleeing Kuwait and Iraq in mid-1992. Recently it played an important role as a partner in organizing the CIS conference.

137. Elsewhere deficiencies have been documented. In Rwanda and Southern Africa a series of shortcomings ranging from over-expenditure to unreliability and misuse of UNHCR-financed assets have been recorded. In some instances (Hong Kong and the Mozambique repatriation) considerable cost savings resulted where UNHCR implemented directly parts of a programme otherwise undertaken by IOM. Criticism within UNHCR of IOM has been further fuelled by a 1991 decision to waive some reporting requirements and questions raised in some situations about IOM's neutrality.

138. Defenders of IOM have suggested that criticism is a reflection of UNHCR's institutional arrogance, and that UNHCR suffers many of the same shortcomings. Given the size of the budget spent through IOM, a detailed review of operational experience to date would appear appropriate and timely and could serve to clarify these differences.

Self-implementation

139. The degree to which UNHCR should itself implement projects directly rather than through partners is an old but still ongoing debate. The Statute has been interpreted to discourage it, although it does not do so explicitly. High Commissioners Schnyder and Hartling were opposed to implementation by HCR, while Sadruddin, Hocké and subsequent High Commissioners favoured it under certain circumstances. In general, policy has been to discourage it, while operational necessity has always required an element of it.

140. Current policy, tacitly reaffirmed by Delphi, is contained in paragraph 1.1 of Section 5.1 of the Manual. It states that direct implementation should be the exception and only occur:

• when there is no viable implementing partner in the country or area;

• at the request of the host Government;

• in the initial stages of an emergency or repatriation operation (until a suitable partner is found);

• when protection or security concerns require the direct operational involvement of UNHCR; and

• for procurement purposes.

141. Practice broadly mirrors the policy. UNHCR now undertakes direct implementation usually only where there are few other options. This was the case with logistics programmes in Namibia and Bosnia. In other cases, in Ukraine for example, UNHCR started implementing through a partner and then switched to direct implementation after the partner was judged inadequate.

142. The arguments in favour of direct implementation are:

• Greater visibility and presence: Direct implementation makes UNHCR more visible. This in turn favours protection, political leverage and fund raising. UNHCR's profile in the Bosnia operation was largely due to the high level of direct implementation.

• Greater control over assets and the programme: Where UNHCR itself implements it has much greater control over a programme and can ensure correct procurement practices and appropriate use of assets.

• Can reduce costs: Particularly in the logistics sector, a number of examples exist where UNHCR has been able to effect significant savings by direct implementation. Contracting others often implies additional overhead costs.

• Enhances experience and skills of organization: It provides valuable hands-on experience to UNHCR staff and enhances the credibility of UNHCR with NGOs and Governments.

143. The arguments put forward against direct implementation are the following:

• It adds a management burden: In addition, it requires sufficient manpower often with technical expertise, which is usually not available.

• In most sectors, NGOs have greater experience and expertise than UNHCR in programme implementation: In Pakistan millions of dollars worth of seed were bought, but nobody knew how to store or distribute them. Shortly before a large portion spoiled, a frantic attempt was made to find an NGO to take them over.

• UNHCR procedures are not geared for self-implementation: Personnel and financial rules are too complicated and bureaucratic. Where direct implementation has been successfully undertaken in the field, it has frequently required the breaking of rules (e.g., on hiring of local consultants, on procurement, and on use of temporary assistance).

• It can prove more expensive than working through local NGOs or international NGOs: This is particularly the case where the latter make a considerable contribution to the programme. The cost of UNHCR international staff is almost always significantly higher than NGO staff.

144. The major constraint to direct implementation is managerial and technical competence. This problem has been overcome through the use of seconded expertise (Nordic and US logistics personnel in Bosnia, for example, or Swiss Disaster Relief personnel in many other programmes) or UNVs. Ignorance on the part of seconded personnel of HCR procedures can often prove an advantage as it prevents these being a restraint to implementation.

145. In recent years, following large procurement and logistics projects in the Gulf, in Bosnia and the Great Lakes region, the level of direct implementation by UNHCR has increased. UNHCR has built up considerable expertise and a comparative advantage in large-scale logistics programmes. This adds to the effectiveness and usefulness of the organization.

146. While many now acknowledge the advantages of direct implementation in some areas, others still oppose it on principle. There are examples where direct implementation has been discouraged by Headquarters although it has resulted in considerable cost saving by the field.

147. While on the whole the criteria for direct implementation set out in the Manual (see paragraph 140 above) seem adequate, it is an obvious weakness that they do not include considerable cost saving as a circumstance when direct implementation would be appropriate.

Commercial Contracts

148. Commercial contracts can be understood as one form of direct implementation. Many staff believe UNHCR should undertake more activities through commercial contracts, especially with local contractors. Commercial contracts offer many advantages compared with other implementing arrangements:

• UNHCR has greater control;

• payment for services or goods can be linked to performance;

• the contract is usually a simple cash transaction with no input of assets, such as vehicles;

• phase-out problems do not occur;

• respective roles are clear;

• contractors can prove more cost-effective than NGOs or Governments; and

• use of local contractors can stimulate the local economy.

149. There are many positive examples of implementation through commercial contracts. In Armenia a shelter programme was contracted out with bids being opened on television. This boosted the local economy and helped to further the practice of competitive bidding in Armenia. In Malawi the logistics programme was managed through a commercial broker. UNHCR arranged low prices without the difficulties of managing a fleet of trucks (in contrast to Ethiopia.)

150. In general, however, the use of commercial contracts is relatively scarce. In most cases sub-agreements are preferred. There are many reasons for this:

• Commercial contractors are often not present or not sufficiently large to be able to handle UNHCR projects;

• Commercial contracts require UNHCR to define exactly what is required. At times UNHCR is not equipped to do this and leaves definition of needs to NGOs or government partners;

• Commercial contracts can be less flexible than sub-agreements and often not appropriate to the changeable environment in which UNHCR works;

• Certain procedures have to be followed prior to the awarding of commercial contracts (unlike with sub-agreements). These procedures create added workload and are often avoided where alternatives exist;

• A degree of technical expertise is required to draft contracts with commercial contractors and to oversee implementation. Technical experts to undertake this task are often lacking;

• It is sometimes suggested that commercial contractors require greater supervision than NGOs as their objectives are commercial rather than humanitarian. There are many examples of contractors taking advantage of UNHCR (although such examples are not limited to contractors); and

• Commercial contracts are subject to much greater scrutiny by auditors than sub-agreements. Auditors are also seen to be less forgiving with non-adherence to procedures.

151. The complications associated with commercial contracts, and the comparative ease of entering into a sub-agreement, lead to some anomalies. In one case in East Africa an international NGO was selected to undertake a borehole project in preference to commercial contractors, although reportedly this added between 5 and 10 per cent to the cost of the project. In another case of rice procurement in South East Asia, Headquarters advised signing a sub-agreement with a supplier after the Field Office failed to present adequate documentation to comply with contracting procedures.

152. One means of overcoming the difficulties involved in drawing up, awarding and supervising the implementation of commercial contracts is to contract through non-governmental or governmental partners. This is frequently undertaken. It has the advantage of taking part of the management burden off UNHCR, but the disadvantage of UNHCR losing an element of control. NGOs also do not always necessarily have the relevant commercial expertise.

153. Lack of control can lead to abuse. Many cases exist where NGOs did not use transparent, competitive bidding procedures to award contracts. There have been recent examples of irregularities with a well-known French NGO and a large US NGO. Cases have also arisen where UNHCR has used NGOs for contracting, usually where there is great urgency, precisely because they in practice are not obliged to adhere to the same strict procedures as UNHCR.

154. Commercial contracts are more suitable for some activities than for others. They work well in a broadly stable environment where the final output of a project can easily be measured (e.g., wells, school construction, demining, and logistics). They work less well where community involvement is required to ensure sustainability (e.g., pump maintenance) or where the output cannot be measured so easily (as with community services).

155. While the changeable and remote circumstances in which UNHCR often works and the crucial roles of Governments and NGOs will always set objective limitations to the use of commercial contracts, it is generally believed that there is room to use them more. Various measures would encourage this, including:

• a revision of rules and procedures regarding commercial contracts, including a review to determine why they are so often side-stepped or not adhered to in the field; and

• a more active role of UNHCR in defining needs - as is intended by Delphi - and increased use of technical expertise (see paragraphs 328-331 in Part II) to this end.

Military Forces

156. The use of military assets in emergencies, in particular national armies assisting in providing relief after natural disasters or refugee influxes, has a long history. The first large-scale example of a foreign or international force being given the primary task of assisting in providing humanitarian aid was Operation Provide Comfort during the 1991 Gulf emergency.

157. Operation Provide Comfort initiated a series of examples of close co-operation between UNHCR and international military forces. In Bosnia, a peacekeeping force provided security, logistics and engineering support without which the operation could not have been carried out. In the Great Lakes emergency, the airlift cell, made up of representatives of different air forces and established for the Sarajevo airlift, had its tasks expanded and its member Governments provided airlift and logistics support throughout the area. Subsequently UNHCR sought military support from Governments in other areas through "Service Packages."

158. In a 1995 concept paper prepared for ExCom members, Service Packages were defined as "self-contained facilities and services mostly but not exclusively from militaries, in sectors where HCR capacity was overwhelmed." The term was coined when the UNHCR emergency response capacity was stretched beyond its limit by the Goma influx and UNHCR sought self-supporting units capable of implementing specific tasks under loose overall UNHCR co-ordination.

159. The main advantages of working with the military are:

• Manpower and skills: Military forces have a very large body of manpower with a great variety of skills (including logistics, road and bridge building, site preparation, health care, emergency medical care, and food distribution) which can be used for humanitarian ends.

• Resources: The military are - certainly when compared to humanitarian organizations - usually over-resourced. This applies especially to logistics (e.g., trucks, light vehicles, loading equipment, and earth-moving equipment). What the US Army spent on fuel alone during Operation Provide Comfort was significantly more than the total annual budget of UNHCR.

• Organizational capacity and discipline: The discipline and organizational capacity of the military are particularly valuable in emergencies, where disorganization and indiscipline costs lives. In Bosnia, the UNHCR organizational culture benefited considerably from the approach of seconded military officers.

• Protection: In certain situations, the protection provided by military escorts can prove essential.

160. The disadvantages of working with the military in summary are:

• Humanitarian integrity: Questions have been raised about the appropriateness of using arms- bearing soldiers to carry out humanitarian tasks. Use of the military can either compromise, or lead to misconceptions about, the neutral, non-political nature of an operation.

• Political constraints: In many situations Governments are reluctant to accept foreign military forces within their territory regardless of the nature of the tasks they carry out. The military are not trained to be sensitive to political concerns.

• Cost: While this may not be charged to aid budgets, the military are expensive as partners. After the Great Lakes emergency, it was estimated that the military airlift cost 4 to 8 times more than its civilian equivalent.

• Security: The military and the Governments behind them are usually more security conscious with soldiers than with humanitarian personnel, as harm to their soldiers on foreign soil has far- reaching political implications. In Bosnia this led to the paradoxical situation that unarmed, humanitarian personnel regularly exposed themselves to greater risk than the military.

161. The pros and cons of working with the military have been much discussed in different fora. UNHCR's 1995 concept paper and the recent joint evaluation of the Rwanda emergency detailed the constraints of Service Packages. One element highlighted by the evaluation was that in the desire to cover gaps through military capacities, there was the tendency to overlook the skills provided by NGOs.

162. After initial enthusiasm, there is now a growing consensus, reflected in inter-agency guidelines, that the use of foreign or international armed forces to assist directly in humanitarian emergencies should be restricted to those extraordinary cases where the response of traditional organizations proves insufficient. Logistics and airlift capacity of the military have broader application. More opportunities could also be sought to learn more from the organizational skills of the military.

CURRENT ISSUES

Auditor and Donor Concerns

163. As indicated in paragraph 48, over the past five years an increasing number of audit observations has focused on UNHCR implementing partners. The criticisms of the auditors are frequently echoed by the ACABQ (a highly influential UN advisory committee) and by donor Governments.

164. Over the past five years an estimated 50 per cent of observations in audit reports undertaken by the internal auditors (UNHCR Section of the Audit and Management Consulting Division of the United Nations), and approximately 30 per cent of recommendations in audits undertaken by the External Auditors, relate to non-compliance by implementing partners with the terms of implementing agreements and insufficient control and monitoring by UNHCR.

165. The following are the main recurring weaknesses of implementing partners highlighted by the internal and external auditors:

• inappropriate and unauthorized disbursements, overpayments and improper use of UNHCR funds;

• aberrant procurement practices;

• inadequate accounts or deficient accounting systems (i.e., no separate bank accounts or bank reconciliations, incomplete accounting records, no control of purchases);

• lack of record or adequate control over UNHCR property, especially vehicles; and

• inadequate reporting and delayed submission of reports (SPMRs).

166. The following, related shortcomings of UNHCR have also consistently been highlighted by the auditors and other studies such as the 1992 CIDA Universalia report:

• lack of adequate financial control of partners;

• insufficient project monitoring;

• lengthy delays in signing agreements and issuing LOIs;

• non-enforcement of the audit certification clause.

167. To address these issues, in 1994 DCMS commissioned a study on financial management of implementing partners. The study made the following recommendations:

• the introduction of financial pre-qualification requirements for partners;

• the undertaking of selective interim audits of partners;

• expansion of the functions of Finance Officers to include monitoring of partners;

• simplification of the Programme Management Cycle; and

• more administrative and finance training of partners and the issuance of 'partner friendly' explanations of HCR procedures.

While at the time the study was largely ignored, one and a half years later most of these recommendations re-emerged.

168. The recommendations of the DCMS study regarding financial pre-qualification requirements were explicitly rejected by a SCAF paper entitled "UNHCR and its Implementing partners" (EC/1995/SC.2/CRP.27, 4 September 95). The paper emphasized the adequacy of current selection procedures and stated: "With regard to selection, experience suggests that the criteria set out in the Annex (i.e., those of the UNHCR Manual) are more appropriate than 'pre-qualification' or other assessments independent of the specific situation."

169. From mid- to late 1995 there was a growing realization that this stance was untenable and changes were required. A series of IOM/FOMs (including FOM/97/95 and FOM/88/96) and a SCAF paper (see paragraph 174 below) drew attention to the auditors' criticisms and announced initiatives to correct them. The IOM/FOMs linked poor management of partners to poor selection and monitoring by UNHCR. To correct this, the IOM/FOMs made proposals - which grew more concrete from IOM/FOM to IOM/FOM - to undertake a pre-assessment of the financial management capacity of partners prior to selection. It was now acknowledged that current selection procedures were inadequate and required strengthening.

170. The 1995 auditors' report, issued in draft in June 1996, had brought the issue to a head. The auditors for the first time threatened to qualify the accounts. The report reiterated many of the familiar weaknesses (see paragraphs 165-166 above) and made the following recommendations on implementing arrangements, stating that UNHCR should::

• ensure regular independent audit of its partners' accounts and should obtain audit certificates;

• introduce precise and well-defined performance indicators in sub-project agreements and defined time schedules for implementation;

• undertake an assessment of capabilities of partners to implement major projects;

• ensure that SPMRs are received regularly and used for project monitoring; and

• review procurement procedures of partners.

171. The auditors' calls for enhanced monitoring and assessment of partners were echoed by the ACABQ (A/AC.96/865. ADD 4), which also highlighted the inadequacy of UNHCR training in financial management for partners. In the autumn sessions of the Standing Committee and at the 1996 ExCom meeting the same points were taken up. The need for improvement had become imperative.

Audit Certification

172. The first recommendation in the Report of the Board of Auditors for 1995 was repeated from the 1994 report and referred to non-compliance of partners with the audit certification requirement. This issue was given prominence, because for the auditors, audit certification represents the most basic and definitive means of demonstrating financial management capacity and certifying the appropriate and correct use of UNHCR funds. 173. The standard clause in sub-agreements, requiring implementing partners to provide audit certificates within six months of the final date of project closure, had proved very hard to enforce. Although a larger number must have been received, in 1995 the desks at Headquarters had recorded the receipt of audit certificates for less than three per cent of all projects.

174. To address the problem a paper was prepared for the Standing Committee entitled "Audit of Implementing Partners (EC/46/SC/CRP.45 of 16 July 1996)". The paper argues that audit certificates provided six months after project closure were "too late for corrective action". Instead it suggests making it a pre-qualification requirement for working with UNHCR that all non-governmental partners present their audited annual accounts, to allow UNHCR to assess their financial management capacity in advance. In addition, UNHCR would subsequently ensure an independent audit of 15 -20 percent of partners annually.

175. The paper met with a mixed response. The auditors, the ACABQ and some Governments welcomed the notion of enhancing the selection procedure through the introduction of a pre-qualification requirement. However, the principle that this could substitute or diminish the need for audit certification was not accepted.

176. The external auditors, in a position broadly supported by the ACABQ, insisted that anything less than 100 per cent audit certification would require them to qualify the accounts. The qualification of the accounts, although not unprecedented - it has allegedly happened before to both UNDP and UNICEF - could have serious fund raising implications.

177. The outcome of the audit certification issue will depend on the extent to which the external auditors can be convinced to accept diminished audit coverage or the use of annual audited accounts in lieu of audit certificates for individual projects. It could also depend on donor willingness to finance the associated costs. The cost of auditing all UNHCR-financed projects has been estimated at over 10 million dollars.

178. The linkage of the audit certification issue with selection procedure was not accepted by the auditors. As indicated in paragraphs 169-171, regardless of the outcome of the audit certification issue, the need to enhance selection procedures with possible pre-qualification requirements remains. The merits of different proposals for enhancing selection procedures are examined in Part II.

Delphi and Implementing Partners

179. The scattered references in Delphi to implementing arrangements indicate similar conclusions to those emanating from the audit reports. Control of implementing partners is seen to be lacking, and a more reflective attitude to selection required.

180. Some 40 per cent of Delphi cells made comments on implementing partner arrangements (by comparison almost all made comments on human resource issues). The following are some of the recurring suggestions, listed in order of frequency. There should be:

• greater control of partners;

• better information flow to and more training of partners;

• simplified sub-agreements, with more budget flexibility and with performance criteria;

• clarification of the audit certification and overhead issues;

• more sub-agreements with local NGOs;

• a database on implementing partners; and

• greater accountability in the selection of partners.

181. The main Delphi documents foresee greater flexibility at field level in determining the content, timing and duration of implementing arrangements. This is combined with a more structured approach to selection and greater accountability of the Field Office for the performance of partners.

182. A more structured selection process is conceived of as part of a more structured approach to overall operations planning. The aim is to ensure that implementing arrangements are determined by the needs and objectives of the operation rather than the other way around. The main tool for achieving this is the Operations Plan.

183. In the formulation of the Operations Plan consideration will be given to what should be implemented by UNHCR, by local agencies and by international organizations. Phase-out potential and contribution to durable solutions will be taken into account in selection. Evaluation of partners according to predetermined criteria reflecting the objectives of the programme is also foreseen.

The Overheads Issue

184. In 1996 the long-standing issue of UNHCR financing of international NGO Headquarters' overheads was resolved. Previous policy had been not to pay the Headquarters' overhead costs except on an exceptional basis, which had to be negotiated case by case in the field. This had led to large discrepancies in the rates paid to different agencies and caused friction with some NGOs whenever UNHCR tried to enforce the original policy of not paying any Headquarters' overheads. Regularization was considered overdue.

185. A uniform policy was adopted at the 1996 Standing Committee meetings of 10-11 April and of 26-27 June. The result was a compromise. It was aimed on the one hand at answering the concern of many international NGOs of using a disproportionate amount of funds raised from the public to pay for Headquarters' support costs for UNHCR projects. On the other hand it was intended to discourage the unhealthy situation where an international NGO had come largely to depend on UNHCR to be able to fund the activities of its Headquarters.

186. Under the formula adopted, UNHCR payment of an NGO's Headquarters' overheads is linked to the NGO contributions to the project. The solution states that UNHCR will pay up to a figure of five per cent of total operational costs towards the NGO's Headquarters' costs, where NGO contributions to the project - defined as costs which UNHCR would otherwise have to incur - match the overall contribution of UNHCR to all overhead costs including those incurred in country.

187. The weaknesses of the formula are the fact that it is complicated, and it depends on an evaluation of the NGO contribution to a project, something which is notoriously difficult. The formula is complicated to apply because it brings in many variables: UNHCR's contribution to the NGO's in-country support costs, the NGO's previous year's contributions, and deduction of those elements from the project which do not require Headquarters' support.

188. A simplified formula would be preferable. Complications could facilitate misunderstanding and could encourage manipulation of accounts. It is believed that the slightly higher cost to UNHCR of a simpler formula would be more than made up for in the hours saved in negotiating, explaining and monitoring the more complicated formula. A simplified formula could take the form of the following:

UNHCR will pay up to a figure of five percent of the total value of a project towards an international NGO's Headquarters' support costs where the agency is able to demonstrate it is making an equivalent contribution to the operational cost of the project.

Conclusion

189. Delphi, recent IOM/FOMs, the auditors and Governments concur in the need to develop a more structured approach to implementing partner selection and monitoring. Three basic means of achieving this reoccur:

• introduce a formal selection procedure with a possible pre-qualification exercise;

• undertake assessment of partners according to pre-established criteria; and

• undertake more training of partners, in particular in financial management.

These topics are discussed further in the second part of this review.

PART II: IMPROVING SELECTION

WEAKNESSES IN CURRENT SELECTION PROCEDURES

190. The quality of service UNHCR provides to refugees and other beneficiaries depends largely on its implementing partners. They undertake the vast majority of UNHCR assistance programmes and an important part of UNHCR protection activities. Approximately US$ 500 million is disbursed annually by UNHCR through partners, which in turn contribute hundreds of millions of dollars more in cash and in-kind resources to UNHCR beneficiaries.

191. While implementing arrangements determine the effectiveness of a programme, they do not always receive the attention they would merit. Donors, auditors, Delphi and to some extent our partners themselves coincide in the need for UNHCR to adopt a more reflective approach to implementing arrangements, in particular to implementing partner selection (see paragraphs 163-183 in Part 1).

192. There is seldom formal discussion on implementation strategy at the outset of an operation or an attempt to determine implementation policy on a regional or country basis. In former Yugoslavia, for example, implementation arrangements varied from Republic to Republic for reasons that were not always clear.

193. Implementing partner selection is decentralised. This causes a discrepancy in levels of accountability for administrative and programme expenditure. By signing a sub-agreement, a UNHCR Representative may hand over millions of dollars to whomever he/she chooses if the amount concerned is within the approved budget, but the same Representative may not order stationery, purchase a fax or rent an office without consultation and approval from Headquarters.

194. The approach to implementing partner selection can be ad hoc. Unlike for commercial contractors, there is no elaborated selection procedure. Section 5.1 of the UNHCR Manual contains two pages of loose guidelines (see paragraph 254 below) but no mechanism exists to ensure they are applied. While in some programmes (Sudan in 1985, Mozambique in 1992) they have served as a basis for more elaborate selection criteria, they are not usually consulted. In the absence of procedures there is scope for cronyism, abuse and fraud.

195. The criteria applied by UNHCR staff in selection can vary. Prior experience of the staff member and personal rapport with the agency representative can make a difference. Most believe there should be a mixture between government and non-governmental agencies. Beyond this some favour local NGOs, while others prefer international agencies. Some are particularly sensitive to donor preferences, while others believe as much as possible should be done through Governments. Some take cost-effectiveness into account in selection while others, if funds are available, do not.

196. Rarely are attempts made to gather information prior to the appointment of new governmental or non-governmental partners. This has led in a number of cases to agencies which have performed poorly in one operation, becoming partners again in another.

197. Selection is often a passive rather than active process. UNHCR at times waits for partners to come to it with project proposals or simply chooses those which are present and active, rather than seeking out partners to meet predefined needs. Agencies can be appointed on a 'first come, first serve basis', particularly when UNHCR is under pressure to show fast results.

198. Although in most situations there is reasoning behind the selection of one partner over another, the reasoning is almost never documented. The absence of a transparent, pre-established procedure and justification requirement makes UNHCR vulnerable to questions and criticism from donors and partners. UNHCR recently, for example, had difficulty justifying to European donors and other agencies the signing of sub-agreements for some US$ 25 million with UMCOR, an American Methodist NGO.

199. The long-term implications are not always borne in mind when partners are first selected. Once a partner has been selected it is hard to change it. Change can be contentious as well as expensive, in particular where a large amount of non-expendable property has been handed over.

200. Inadequate attention to implementing partner selection has led to a series of examples of inappropriate choices and poor performance. In Mozambique a health NGO was selected to dig wells, and in Uvira an agricultural NGO was chosen to do community services. In Somalia an agency was paid for digging boreholes which did not exist.

201. Two main justifications are put forward for not paying more attention to implementing partner selection: lack of time and manpower, especially in emergencies, and lack of choice. The staff and time dedicated to the selection process can be a reflection of the priority it is given (this paper argues it should be given higher priority). The degree to which UNHCR has a choice in selecting partners is examined in the following section.

THE ISSUE OF CHOICE

Choice with Governments

202. Opinions differ, sometimes within the same programme, on the leeway UNHCR has with Governments to determine implementing arrangements and influence government structures. Some believe UNHCR overestimates its influence, while others think that UNHCR has more influence than it realises.

203. In practice, the level of UNHCR influence varies considerably from operation to operation. In some countries such as Malawi, Benin and Armenia, UNHCR has had considerable influence. In other programmes, such as those in Iran, Mexico or China, there has been little or no choice.

204. Various factors play a role in determining the level of influence UNHCR can have with Governments. Some of these are:

• Political/security concerns and perceptions regarding sovereignty: Where refugees are seen to affect national security, as was the case in Pakistan, or where issues of sovereignty are an overriding concern, UNHCR is likely to have little influence.

• Administrative structure and tradition: The degree to which government structures are centralised, their level of development and the attitude of the Government towards NGOs will influence the structures to emerge and UNHCR's ability to be of influence.

• Early arrival of UNHCR: Where UNHCR arrives early, before government structures to deal with refugees have been established or consolidated, UNHCR has more possibilities to influence structure. This was the case in Benin and Georgia.

• The degree to which UNHCR presence was sought after by the Government: In cases where there has been some resistance from within the Government to the establishment of a UNHCR programme, such as in Iran and Mexico, UNHCR is left with little influence. Where UNHCR presence is seen as crucial by the Government, as was the case in Malawi, UNHCR has greater influence.

• Senior level access: High-level access can give UNHCR greater influence in determining implementing arrangements.

• UNHCR credibility: Where UNHCR is highly regarded, it has greater leeway. In Central America, where UNHCR's reputation was strong, it had a lot of influence in determining implementing arrangements despite the political and security sensitivities associated with refugee issues in the region.

205. Many believe, in the case of large programmes, that UNHCR could make more use of the budget as leverage to influence implementing arrangements. Large budgets can restrict choice. Examples exist of small programmes where UNHCR had a high degree of influence (Tajikistan, Armenia) and of large programmes where UNHCR influence was restricted (Ethiopia, Sudan). In some cases Governments were unconcerned about implementing arrangements until the resource implications became apparent.

206. A further factor affecting choice with Governments is time. The influence UNHCR could have through the programme budget is often diminished due to pressure to spend quickly. In Pakistan, for example, some believed that only a large special structure could absorb and disburse the large amounts of resources in the limited time the donors expected and the refugees required.

207. The degree of awareness of UNHCR Representatives of the importance of implementing arrangements can also be a crucial factor. Large special structures have been created unwittingly, growing over time out of small co-ordinating bodies. In other cases, by contrast, such as in Malawi or Benin, a conscious and successful effort was made to apply lessons learned from elsewhere and to create 'ideal' implementing arrangements.

Choice with NGOs

208. The level of choice UNHCR has with NGOs is basically determined by three factors:

• the number of NGOs present;

• the degree to which the NGOs present wish to work with UNHCR; and

• the pressure brought to bear by host or donor Governments to work with or avoid working through individual agencies.

209. The extent to which local NGOs are present varies considerably and does not always follow a clear pattern. Some countries have strong NGO traditions, while in others, although possibly close by, there are few if any NGOs. In Guinea and Benin, for example, there were few NGOs while in Ghana and Ivory Coast there are many.

210. A factor that can influence the number of international NGOs present is the profile of an emergency. Large media and NGO presence went hand in hand in Somalia, Northern Iraq and the Great Lakes region. Prior to the Goma influx it had proven hard to attract international NGOs to Zaire. The profile of activities and of a refugee site can also be determining factors; NGOs were abundant in Ngara and Goma but scarce in Uvira. Many more NGOs work in health or with unaccompanied minors, than in sanitation or site planning.

211. In some programmes although NGOs are present, UNHCR is not always able to work with them. In Russia some of the NGOs present were unwilling to work with refugees. In other cases NGOs had sufficient funding from other sources and were unwilling to work with UNHCR. This occurred initially with MSF France in the Chechnya emergency and with other NGOs in Bosnia.

212. Where NGOs are more dependent on UNHCR either for funding or for administrative purposes, UNHCR has more choice in selection. In Tanzania, UNHCR was in a privileged position as it participated in the government screening process. Elsewhere in the Great Lakes region UNHCR had a large range of partners to choose from as ECHO funds for NGOs were channelled through it.

213. The degree to which UNHCR can work with NGOs is also dependent on the attitude of the host Government. Governments sometimes prohibit or curtail the involvement of NGOs, in particular where the refugees are located in insecure areas or are considered a potential security threat (initially in Pakistan and Mexico). In other cases the Government may designate a specific NGO to work with UNHCR (in Malaysia for example).

214. Donor Governments also play an increasingly important role in restricting the choices UNHCR has with NGOs. There are various ways this is done: individual donor Governments can earmark funding or in-kind contributions through UNHCR for their national agency (which occurs typically with Germany and GTZ); pressure can be brought to bear by donor representatives in the selection process on UNHCR in the field and at Headquarters; and Governments can provide the NGO concerned with sufficient resources to make a large contribution to a project at the outset, thus obliging UNHCR to take it on as a partner (see paragraph 261 below).

215. Where NGOs are imposed on UNHCR this can add significantly to the cost of a programme. UNHCR loses leverage in negotiating budgets and can be obliged to accept arrangements which are not cost-effective. An example occurred in the Bosnia shelter programme. UNHCR efforts to implement all through a single agency, and thus achieve economies of scale, foundered when the opposition of a major donor became apparent as the NGO was from elsewhere. At a significantly greater cost, UNHCR was compelled to work through a number of agencies from other countries.

216. UNHCR is often in a poor position to resist donor pressure. In some circumstances (see paragraph 272 below) an imposed arrangement can be the only way to gain funding for a programme. In normal circumstances, however, UNHCR would be in a stronger position to oppose donor pressure if it consistently used a transparent selection procedure.

Increasing Options

217. Implementing arrangements are often regarded with excessive fatalism. While there are some factors determining the level of choice over which UNHCR has no control, there are other factors influencing choice which depend on UNHCR. Where UNHCR arrives early, researches local implementing options, discusses implementing arrangements at the highest government level, and has a credible programme, it is usually able to have some influence on government and NGO implementing structures.

218. The level of choice in implementing arrangements can to some extent be a reflection of the time and effort dedicated to seeking and creating options at the outset of an operation. The first assessment missions should include implementing arrangements within their Terms of Reference. Questions could include:

• What local structures are available to implement the programme ?

• Do they have the capacity to undertake the tasks required and expand if necessary ? In what areas would strengthening be required ?

• What is the position of the Government on implementing arrangements ?

• What should be the relative distribution of tasks between Government and NGOs ?

• What international NGOs are present ?

• Is it necessary to try to encourage new agencies to come or to stimulate the creation of agencies ?

• What will be the long-term effect of the implementing arrangements under consideration ?

• What arrangements are most likely to contribute to the long-term objectives of the programme ?

219. Reflection and research on implementing arrangements should form an important part of contingency planning. The current contingency planning guidelines could be elaborated in this respect. They contain little on implementing partner selection procedures and restrict themselves to highlighting the work of five NGOs, IOM and some UN agencies. Under the approach they foresee, agencies present in a given situation state what they will be able to contribute should an emergency arise. This approach is of limited use as a reliable basis for selection.

220. When a programme has been running for a number of years and implementing arrangements are already defined, options seem more limited. Periodic reflection on implementing arrangements may still be appropriate. Questions which can be raised in such a review include:

• Are current implementing arrangements optimal ?

• To what extent could shortcomings be overcome by changing partners ?

• Are other options available locally ?

• Could other options be created (bringing agencies in, strengthening and expanding the functions of some agencies, undertaking direct implementation) ?

• What are the political and financial costs of change ?

• Over what period would this added cost pay for itself ?

221. While there are always constraints, in many situations there is more choice than is apparent. Government structures are seldom cast in stone and UNHCR can sometimes influence their evolution over time. In the case of some large structures, UNHCR has been able to achieve improvements over time through seeking donor support and by exercising firm and persistent pressure.

222. Where UNHCR has absolutely no control over implementing arrangements, the question may be relevant as to whether the programme should continue. Some organisations such as the ICRC will not work where they have no influence or control over implementing arrangements. UNHCR has accepted to work in situations where implementing arrangements were imposed and where these were abusive. Stricter policy in this regard could eventually also give UNHCR more choices.

223. A number of measures could be taken to enhance the choice UNHCR has with NGOs. Where NGOs are lacking more could be done by Headquarters to try and solicit international NGOs to come to a country. In Ngara, CARE came within 24 hours of being approached by Headquarters.

224. Cost can be a constraint to bringing in outside agencies, particularly as UNHCR pays a limited part of NGO Headquarters' overheads. Another constraint is that UNHCR is sometimes uncertain of its own commitment to an operation (in Chechnya, for example, the operation was initially extended almost by the month), and is thus in a weak position to convince others.

225. At present NGO liaison functions focus mainly on UNHCR's traditional partners. More efforts could be made to investigate potential new partners, in particular for those activities where partners are usually lacking (e.g., sanitation, site preparation, logistics, and income generation schemes). Efforts could also be made to learn more about agencies less familiar to UNHCR, in particular Islamic NGOs.

226. Where choice with NGOs is restricted by host government policies, UNHCR could also play a more active role in lobbying on behalf of NGOs. UNHCR sometimes shies away from involvement in disagreements between NGOs and host Governments. UNHCR could strive more consistently to advocate for NGOs with Governments and assist Governments when they draw up legislation and administrative procedures regarding NGOs.

IMPROVING SELECTION

Elements of a Selection Process

227. As indicated in paragraphs 163-183 of Part I and 190-201 above, there is a broad consensus on the need to develop a more structured approach to selection of implementing partners. This implies the introduction of a transparent selection procedure which would depersonalise selection and ensure more forethought in the process. The introduction of a procedure should not unduly reduce flexibility or add delays.

228. A number of proposals have been put forward as elements in an improved selection procedure. The most important components argued for in this report are:

• Ensure reflection on implementing arrangements is an integral part of all operations planning; This is a theme of Delphi (see paragraphs 179-183 in Part 1) and a key conclusion of this report (see paragraphs 218-220 above). To encourage reflection, a section could be added to the Country Operations Plan or its Delphi successor under which Field Offices would be required to explain and justify the implementing arrangements adopted.

• Pre-assess agencies, in particular their financial management and technical capacities prior to engaging them: This is a reoccurring recommendation of audit reports (see paragraphs 163-171 in Part 1). As a means of ensuring pre-assessment, a formal pre-qualification exercise was proposed by an earlier review and advocated by a recent Standing Committee paper. Pre-assessment is examined further in paragraphs 233-240 below.

• Formulate more elaborate criteria to assess and select and ensure their application: The criteria in the Manual are useful, but not comprehensive. Other possible criteria are examined in paragraphs 245-272 below. Such criteria can be applied through questionnaires adapted to the specific field situation.

• Involve more staff at the Field Office level in selection: Selection is usually left to Programme Officers. UNDP and UNICEF by contrast have selection committees. Many believe that UNHCR technical specialists and Finance Officers should be more systematically involved in selection. This is examined in paragraphs 324-334 below.

• Involve partners in the selection process: While selection remains the ultimate responsibility of UNHCR, the process is more likely to gain the co-operation of partners, if they are consulted on criteria and procedure.

Tendering Projects

229. The idea has also been put forward that UNHCR should systematically put projects out for tender. While international advertising is a requirement for commercial contracts over US$ 20,000, multi-million dollar sub-agreements can be signed without more than one agency being consulted.

230. Tendering with NGOs is resisted both within UNHCR and by NGOs. It is considered inappropriate for the following reasons: the relationship with NGOs is not commercial as they usually contribute to the project; tendering is contradictory to a relationship of partnership; it encourages competition in an environment where co-operation is more appropriate; it restricts flexibility in programme implementation; and it presupposes a broad choice of NGOs, which in practice is often lacking.

231. While in most situations a formal selection procedure, which among other elements takes cost into account, is preferable to tendering, in certain situations tendering could be appropriate. In Armenia something approaching a tendering process was used with success to identify partners to undertake quick impact projects (QIPs).

232. The circumstances in which tendering of projects could usefully be pursued are the following: where the contribution of international NGOs to projects is negligible; where there are many international NGOs present; where UNHCR has difficulty applying its own selection criteria due to donor pressure; and where projects can be clearly defined. In Bosnia tendering projects from 1994 onwards in certain sectors would possibly have saved millions of dollars and allowed UNHCR to demonstrate a clear rationale for selection to donors and agencies.

Pre-assessment

233. At present efforts to assess new partners prior to selection are inconsistent. Agencies, whether local or international, are not systematically subjected to scrutiny prior to appointment. Requests for information, for example, on experience with an agency elsewhere or the nature of its legal status, are rare. When such information is sought from Headquarters - which was the case with the NGO British Direct Aid in Rwanda - the response can be slow.

234. While there is a need for more pre-assessment of potential partners, the introduction of a formal pre-qualification exercise as suggested by a recent Standing Committee paper (see paragraph 174 in Part I) is not considered appropriate. It implies a sub-contractual relationship, rather than one of partnership. If undertaken independently of a selection procedure, it would also add an extra work burden and potential cause of delay.

235. An alternative proposal builds on the procedure contained in the UNHCR Manual and the approach indicated in recent IOM/FOMs (see paragraph 169 of Part 1). In the Manual pre-assessment of partners is undertaken as an integral part of selection procedure.

236. The Manual also differentiates between two types of criteria. Section 5.1 (paragraphs 3.2 and 3.3) of Chapter 4 of the Manual lists four basic conditions agencies are required to meet to be considered as UNHCR implementing partners. Nine additional criteria are listed (see paragraph 254 below) which should also be taken into account in assessing partners prior to selection, but which are not mandatory prerequisites for working with UNHCR.

237. Under the procedure proposed, potential partners would be assessed and selected on the basis of the criteria contained in the Manual and those proposed in paragraphs 245-272 below. The criteria could be applied through a checklist or questionnaire. This could be drawn up by Field Offices, which in the process would adapt the criteria to their particular needs and circumstances. The use of such a checklist or questionnaire could be made obligatory.

238. The stringency with which the criteria would be applied would vary according to the size of the project and the choices at hand. Partners for large projects, in particular where a number of choices were at hand, would be subjected to greater scrutiny than those selected for small projects.

239. While the application of some criteria would vary according to circumstance and the size of the project, others would be applied strictly regardless of circumstance. They would constitute minimum, mandatory prerequisites for working with UNHCR. The application of minimum, mandatory prerequisites along the lines of those proposed below obviates the need for an additional pre-qualification exercise.

240. Where an agency was unable to meet the mandatory criteria but was still considered valuable as a partner, the criteria would indicate where strengthening was required. In the case of non-compliance with the minimum mandatory prerequisites there would be an obligation for the agency to declare its willingness to receive guidance and strengthening from UNHCR to correct this, and an equal obligation for the UNHCR office concerned to commit itself through training or budgetary provision to undertake the task. A time-frame for achieving this could also be set.

Selection Criteria

241. Those against the use of selection criteria suggest that broad guidelines are sufficient and 'stricter' criteria not appropriate to the situations UNHCR works in and to the limited choices usually available. Interviewing the individuals of the agency concerned is more important than the application of abstract criteria.

242. The advantages of using criteria are believed to outweigh the disadvantages. They give UNHCR a rationale for working with some and rejecting other partners, and provide UNHCR with a means of resisting the imposition of agencies. Where partners do not meet the criteria but there are no other choices, criteria give an indication of where strengthening is required. The World Bank, WFP, UNICEF, UNESCO, UNFPA and funding NGOs such as Oxfam, have all elaborated detailed selection criteria.

243. The formulation of final criteria - should the recommendations of this report be accepted - would ideally be undertaken together with partners. ICVA has also suggested that there should be a recourse procedure for NGOs that are excluded. The approval by Governments of the criteria, through the Executive Committee, would give them added authority.

244. For a selection procedure to be credible it has to be universally and consistently applied. This implies formulating criteria which will allow for this by taking account of the constraints that exist in practice. To ensure that the procedure is always applied, a documentation requirement could be introduced when selecting partners for projects over a certain amount (say US$ 500,000).

Minimum Criteria: Legality, Humanitarian Integrity and Financial Management Capacity

245. Legal status: As one of the four minimum conditions for working with UNHCR, the Manual includes the prerequisite that an agency be legally registered where they are operating. It also requires them to "have established a working relationship with the Government". However, in some situations there are no formal registration procedures, nor a recognised Government.

246. In order to take these cases into account, while still requiring agencies to gain official sanction, this prerequisite could be modified as follows:

"UNHCR partner's shall be constituted and work in accordance with the legal requirements in force of the country/territory concerned and with the agreement of (or alternatively 'without objection from') the relevant authorities."

247. A further prerequisite, listed in the Manual, is the requirement for agencies to be willing to work with all intended beneficiaries and provide assistance on the basis of need only. In practice this prerequisite is often flouted. Governmental partners can be inclined to pursue political ends and NGOs, sometimes created to help specific interest groups, can also be discriminatory. In open conflict situations, it can be unrealistic to expect a local governmental or non-governmental agency to be willing to provide aid to the 'aggressors'.

248. Humanitarian integrity: While UNHCR is sometimes in a weak position to influence the nature of the agency providing assistance, there is room to be more demanding with regard to the manner in which the agency concerned implements projects financed with UNHCR funds. Shifting the emphasis from the nature of the agency to the manner in which it implements UNHCR projects would make it easier for UNHCR to insist on compliance. This criterion should also emphasize - which it does not at present - the requirement that partners be non-commercial, non-profit making. A possible alternative formulation could be:

"UNHCR partners shall be non-profit making, non-commercial entities who neither directly nor indirectly pursue any objectives which are contrary to or which may threaten or undermine the humanitarian integrity of the UNHCR programme. They shall undertake to implement any projects financed with funds provided through UNHCR under the guidance of UNHCR and on the basis of need and without discrimination according to race, religion, ethnic group or any other extraneous criteria."

249. Financial management capacity: A third basic prerequisite listed in the Manual refers to financial management. The Manual requires agencies to open a separate account for UNHCR expenditure and to demonstrate financial reliability through the production of official audit statements. Non-adherence to this has caused many problems with donors, the ACABQ and auditors (see paragraphs 163-171 in Part I).

250. It is frequently stated and there are many examples, that good financial reporting is not necessarily an indicator of good performance. The fact, however, cannot be altered that financial management - perhaps because it is more straightforward to evaluate - is subjected to greater scrutiny than other aspects of a partner's activities.

251. The Manual highlights the importance of a separate bank account. Non-compliance with many other financial requirements foreseen in sub-agreements is equally problematic. This prerequisite could be reformulated as follows:

"UNHCR partners must be able to identify to UNHCR a person, preferably with relevant training, responsible for maintaining the accounts. Agencies should also be able to present accounts which should clearly show income and expenditure, assets and liabilities. Unless there are compelling reasons to prevent this, these accounts should be audited by an independent auditor. The agency should be able to demonstrate its ability to comply with the sub-agreement clauses relating to the need for separate bank account (3.06), reporting and audit (3.08 -3.17) and, where relevant, procurement (7)."

252. The above criteria regarding legal status (paragraph 246) and humanitarian integrity (paragraph 248) should be mandatory prerequisites for working with UNHCR. In the case of financial requirements (paragraph 251) where an agency was unable to meet the criteria, specific measures and a time-frame would have to be agreed upon to provide the relative training and personnel to allow for full compliance (see paragraph 240 above).

Other Criteria

253. A series of other criteria should also be applied in the selection process. These criteria would not be mandatory prerequisites but would have to be applied in any situation where there was more than one agency to choose from. They could also be used in comparing the options of implementation through a partner versus direct implementation or implementation through commercial contract.

254. The Manual lists the following criteria: quality of service; rapid response; local experience; contribution of resources; continuity of staff; UNHCR experience; international or local organisation (compare paragraphs 95-105 in Part I); single NGO or consortium and phase-out potential. These require review and additions.

255. Technical expertise: This is not explicitly highlighted as a criterion by the Manual. The World Bank requires its partners to have a demonstrated expertise and strong track record in the relevant sector. UNHCR by contrast sometimes enters into sub-agreements amounting to hundred of thousands of dollars without prior evaluation of the technical competence of the agency concerned. Greater emphasis needs to be placed on technical competence.

256. To this end the set of minimum technical standards currently being developed by a group of NGOs could be adopted by UNHCR. The role of technical experts in the selection process could also be strengthened (see paragraphs 328-331 below.)

257. Protection: Surprisingly the implications of an implementing arrangement on protection are seldom considered in the selection process. Pre-assessing implementing options and potential partners from a protection point of view could help further protection concerns.

258. Assessment from a protection point of view could be combined with greater efforts to promote protection concerns through additions to sub-agreements. Where UNHCR presence is limited, for example, clauses could be introduced requiring partners to monitor and report on protection concerns.

259. Decisive factors determining the efficacy of an agency for protection purposes are level of commitment and degree of influence. Some also argue that international agencies are preferable for protection to local agencies as they are less vulnerable to local pressure and may report with more distance and objectivity. Local agencies can, however, be more committed and better informed than international agencies. When rooted in the local community they can also be an effective means of promoting protection principles among the local population.

260. NGO contribution: This criterion is included in the Manual and, in connection with the discussion on UNHCR contributions to NGO Headquarters' overheads, has recently received new emphasis. The immense contribution made by NGOs to UNHCR beneficiaries is not always adequately recognised. However, using the contribution of an NGO as a criterion for selection has many limitations.

261. NGO contributions are hard to measure and can be subject to abuse. In some cases NGOs have gained a sub-agreement with UNHCR on the basis of a promised contribution that has never materialised. In other cases NGOs or other agencies have initially made a single, large contribution, thus securing themselves lucrative sub-agreements for many years. Emphasis on a contribution - often made in the form of international staff even when local expertise is available - can also mitigate against working with local agencies. A high NGO contribution need not indicate the most cost-effective means of implementation.

262. Cost effectiveness: Evaluating the overall cost effectiveness of an implementing arrangement, taking into account its cash or in-kind contribution, is considered a better measure for ensuring value for money and for maximising resources for beneficiaries. The cost effectiveness of an implementing arrangement can be established through cost benefit analysis: assessing the relation between the total cost of a project and the value of the resulting benefits.

263. While Programme Officers regularly negotiate on individual budget items, efforts to assess and compare the overall cost effectiveness of different implementing options are less consistent. In the Great Lakes region varying arrangements have evolved in different countries for fleet maintenance and transport. Although in some cases the arrangements have not proved satisfactory, and in all cases the value of the projects concerned is many millions of dollars, no comparative cost benefit studies were undertaken.

264. A tendency to look at cost in terms of individual line items rather than total cost and longer term benefits can lead to waste. Near Jijiga and Hartisheik, for many years UNHCR continued a water tanking operation rather than build a pipeline which in the immediate term would have cost more but in the long term would have saved millions. International health NGOs are sometimes preferred over local health services as they keep better books, but do not necessarily provide better value for money.

265. Some argue that looking at cost effectiveness requires time and skills which Programme Officers, in particular in emergency situations, do not have. Comparing costs, it is sometimes argued, is only meaningful when there is more than one option. This is often not the case. In emergencies, where lives are at risk, cost, it is stated, should not be an issue.

266. These objections apply in a limited number of situations. In many situations there are more options than may immediately be apparent (see paragraphs 202-232 above). The options of direct implementation or implementation through local, commercial contracts are not always considered and rarely costed. The purpose of emphasising cost effectiveness is not simply to reduce costs, but to gain better value for money, which in turn can mean providing a better service to a greater number of beneficiaries.

267. As regards the time available to Programme Officers to undertake cost benefit analysis, a small portion of the savings that would be gained if such analysis were to be systematically undertaken, it is believed, would be more than sufficient to finance the posts to undertake the task.

268. Contribution to local structures: UNHCR recognises the obligation to contribute wherever possible to strengthening local structures. To this end it is UNHCR policy to work in preference through local agencies. In many situations, however, (see constraints outlined in paragraph 98 of Part 1) this is not possible.

269. Where UNHCR is unable to implement through local agencies, the potential contribution of an implementing partner to strengthening local structures should be a selection criterion. A number of indicators can be used to evaluate this:

• the extent to which the agency works through a local partner or assists local agencies or the local authorities;

• the extent to which this co-operation involves training and capacity enhancement of the local partner;

• the extent to which there is a commitment to hand over to a local partner within a set time frame;

• the proportion of local and international staff;

• the degree to which the contribution made by the agency can be sustained by local structures; and

• the degree to which (in particular in repatriation and reintegration projects), the agency abides by national standards and attempts to work in accordance with national development priorities.

270. It has been suggested that UNHCR should only sign sub-agreements with international agencies where there is a commitment to hand over to local agencies within a set time frame (see paragraphs 313 below). Such a requirement could be appropriate in local settlement and reintegration projects, but would be of more limited application in emergency situations, in particular where rapid repatriation was envisaged.

271. In addition to those discussed in previous paragraphs, a number of further criteria have been proposed to assist in implementing partner selection:

• The agency should abide by the Code of Conduct formulated by the ICRC/IFRC and some major NGOs (see paragraph 43 in Part 1): This provides a basic standard in particular for pre-assessing international NGOs (see paragraph 228 above).

• The agency should draw on a real constituency and be rooted in the local community: UNESCO demands that its partners be "representative and close to the communities they seek to serve on the ground." The criterion of legitimacy is of particular importance in assessing local agencies. An indicator of this is the degree of public and government support they enjoy.

• The agency should assist in furthering durable solutions: This is emphasized by Delphi, but in practice can be hard to assess. The degree to which an agency may impede durable solutions, i.e. its phase-out potential, is perhaps easier to judge. This criterion is already contained in the Manual (see paragraph 254).

272. Fundability: A further factor that sometimes has to be taken into account in selection is fundability. In certain circumstances using a less desirable implementing option may be the only way of gaining funding for a project. While every effort should be made to resist donor pressure in selection, the ability of an implementing arrangement to attract funding (in particular funding which would not otherwise have come to UNHCR) should, when a programme lacks funds, be taken into account.

Other factors: Phases of an Operation and Limiting Numbers:

273. A number of other factors affect selection. One important factor is the particular phase of the operation and its objectives; the importance of different selection criteria, and thus the weight they are given in the selection process, will vary according to the stage the operation is in.

274. Emergency phase: In emergencies, it is often argued, there is no time to think of longer-term implications of implementing arrangements or to study all the options available. For this reason the inclusion of reflection on implementing arrangements in contingency planning (see paragraph 219 above) is particularly important.

275. Where no forethought is possible and implementing arrangements have to be conceived of in a hurry, efforts should be made to curtail their long-term implications. Sub-agreements could contain explicit time limitations. They could be subject to review after, for example, four months. Non-expendable property could be loaned with an explicit return date.

276. In emergencies the main selection criterion is speed, the ability to get assistance to refugees as soon as possible. The specialised expertise of international agencies, particularly as local structures are likely to be overwhelmed, is crucial in this respect.

277. Post-emergency phase: In this phase, the priority of working with local partners should predominate. The effectiveness of local partners will be enhanced where they have been involved, possibly in partnership with an international agency, from the initial stages of the operation.

278. Different partners may be required respectively for repatriation and reintegration. The expertise required to undertake repatriation is distinct from what is required to undertake reintegration projects. The logistics required for repatriation operations is often best undertaken by UNHCR directly or by international NGOs who can cross borders with comparative ease.

279. For reintegration projects, UNHCR works ideally with local partners. However, in post-conflict situations in which reintegration operations are typically carried out, local structures are often limited or debilitated. In such circumstances, special efforts (see paragraphs 301-316 below) are required by UNHCR and its international partners to rehabilitate and build up local capacities.

280. The overriding selection criteria for agencies involved in reintegration should be contribution to the local structures and sustainability. They must remain and be able to continue functioning after UNHCR has left. In Nicaragua, the fact that UNHCR's main partner for reintegration, the IRC, left at the same time as UNHCR, did not assist the reintegration process.

281. Limiting numbers: A further factor affecting selection of partners is the issue of numbers. The 1994 audit report recommended that UNHCR should reduce the number of its partners. It has also been suggested that UNHCR should have umbrella agreements with a few large NGOs and wherever possible work with these.

282. There are many advantages to limiting the number of partners involved in a programme: management and control of partners are facilitated, overhead costs are reduced and economies of scale can be achieved. The large agencies which form UNHCR's main partners, where UNHCR works with a limited number of partners, are also more likely to be familiar with UNHCR reporting and programming practices.

283. Reduction of numbers of partners has its limits. The use of a single, predominant partner has led to problems in many situations. The experience with the Malaysian Red Cross, for example, is generally not considered worthy of replication. A monopoly on implementation can lead to complacency and abuse.

284. There are arguments for maintaining a broad variety of partners; a diversity of partners can reflect the diversity among the beneficiaries. Small agencies are often closer and more attuned to the needs of beneficiaries. In some situations there may be a need to maintain a balance between partners (government and non-governmental, or between NGOs of different religious persuasions). Large international agencies, sometimes favoured when numbers are limited, can be more expensive to work with.

285. To facilitate maintaining a variety of partners, while not adding an undue management burden, implementing instruments could vary according to the size of the project. UNICEF uses a simple letter of agreement for grants of less than US$ 10,000. In UNHCR by contrast the same format, including identical reporting requirements, applies whether the sub-agreement is for US$ 20 million or US$ 5,000.

IMPLEMENTING PARTNER EVALUATION

286. Periodic evaluation of the performance of implementing partners has been proposed as a further means, complementary to enhanced selection procedure, of professionalizing UNHCR's approach to implementing partners. It was recommended by the CIDA Universalia Report and the 1994 audit report. The proposal was reiterated in recent IOM/FOMs (see paragraph 169 in Part I) and is advocated by some donor delegations.

287. Different proposals have been put forward of how to evaluate agencies. Some have proposed identifying specific criteria ranging from accounting systems, to technical expertise and grading the agency with a point system according to these. Others have suggested UNHCR should have looser criteria and simply maintain a list of the hundred best agencies, or a blacklist of the worst.

288. There are many constraints to carrying out any form of evaluation of partners. A previous attempt undertaken in the late 1980's failed (see paragraph 30 in Part I). Arguments put forward against evaluation are:

• The environment UNHCR works in is not conducive to evaluation. Situations are too fluid, too changeable;

• Evaluation is not appropriate to a partnership relationship. It can be destructive of this relationship;

• NGOs vary greatly from one operation to another or even within the same operation. Evaluations can be misleading;

• It is hard to have objective criteria. The exercise can become subjective, reflecting the personal relationship of the UNHCR Representative and the agency Representative; and

• UNHCR does not have the skills or the time to undertake evaluations of other agencies.

289. The main arguments put forward in favour of evaluation are:

• It is unprofessional and can encourage abuse to spend hundreds of millions of dollars through partners with no formal system to assess whether this money was well spent;

• UNHCR and agencies can grow complacent working together year after year;

• In dealing with partners UNHCR has a tendency to focus on budget details rather than final output. Evaluation of projects is required to switch the focus to the effectiveness of partners;

• At present evaluation is undertaken informally on the basis of anecdotes. A more formal system would be more objective;

• Some NGOs favour evaluation as it helps to separate the serious from the less serious agencies; and

• Some undertake evaluation and others are evaluated. Oxfam, for example, is compelled by its charter to spend 5-8 per cent on evaluating its projects.

290. The need for improvement on the present system is broadly recognised. Increased evaluation is required to better justify choices, as an incentive to improve performance, and to provide a rationale for reappointing agencies year after year. Its results could serve less to castigate than to indicate areas where strengthening was required.

291. Evaluation can be undertaken on the basis of performance benchmarks (this is proposed in the 1994 and 1995 audit reports) agreed upon with the agency at the outset of the project. This method works particularly well where projects have a tangible output, such as well digging, school building, and food distribution. Additional 'hard' evaluation criteria, which leave little room for subjective judgement, can also be formulated. For example, did the partner report in a timely and accurate fashion ? Did the agency remain within the budget?

292. The partner would be given the opportunity to justify where a force majeure (e.g., new influx, late disbursement by UNHCR, change in programme, or coup d'etat) prevented it from meeting stated goals. As part of the evaluation process, the agency should also be given the opportunity of making suggestions on how UNHCR performance in the specific operation could be enhanced.

293. The level and depth of evaluation should vary according to the size of the project. In the case of most projects evaluation could be a short annual exercise carried out with the co-operation of the agency concerned and leading to a one or two page record of actual performance compared to stated goals and any constraints that arose. In the case of large projects (for example, those over US$ 5 million) a more in-depth assessment led by Headquarters' staff in co-operation with the agency concerned would be required.

294. A more comprehensive evaluation of implementing arrangements could also form part of the Terms of Reference of evaluation and inspection missions. This could be undertaken on the basis of the same criteria as those proposed for selection (paragraphs 245-272). With the notable exception of the 1995 Mozambique evaluation, evaluations of country programmes seldom critically analyse implementing arrangements.

295. Joint evaluations (also proposed by PARinAC in recommendation 122), interagency evaluations, and those bringing together donors and agencies should also be encouraged wherever possible. UNHCR could also seek to share more freely the evaluations it undertakes and encourage other agencies to do likewise.

IMPLEMENTING PARTNER DATABASE

296. Many believe UNHCR should establish an implementing partner database. This was also advocated in a recent JIU study and a July 1996 Standing Committee paper. The database would serve as a reference tool and facilitate responding to frequent internal and external queries on implementing partners.

297. A number of related initiatives already exist. STS maintains vendor notes on suppliers. EPRS has collected information on the emergency response capacity of 35 NGOs. The NGO Liaison Section maintains an NGO directory with basic information on a much larger number of NGOs. DFIS maintains a record of the value and number of projects undertaken by individual partners.

298. Current information tools available at Headquarters on implementing partners are regarded by many as insufficient. The fact that information is disbursed between EPRS, DFIS and the NGO Liaison Section is a disadvantage. The information contained both in the NGO directory and the EPRS database is limited. For example, neither records the type, value and number of projects UNHCR has undertaken with the agency, nor do they give an indication of what agencies excel in which areas.

299. Suggestions on the type of information that could usefully be contained in a database includes the following:

• Name, address and names of principal officers

• Type of organisation (commercial, charitable, governmental)

• Founding date

• Sectoral areas in which they work

• Areas of particular expertise and comparative advantage

• Size of annual budget and main sources of funding

• Number of staff, average contract lengths and staff formation

• Current locations of operation

• Outline of accounting and stock-taking systems used

• Length of deployment time in an emergency

• Year, location and value of projects undertaken with UNHCR

• A record of how the agency has performed in projects financed by UNHCR (this information could be summarised from the evaluations proposed in paragraphs 291-293 above)

• Any other information of note (e.g., record of negative audit comments, or note of extraordinary achievements)

300. To be useful, a database has to be comprehensive, have some level of detail, and be up to date. To achieve this requires considerable staff resources. The burden of collecting information and maintaining it could be shared between Headquarters and the field. Assistance could also be sought from an NGO umbrella organisation, such as the International Council of Voluntary Agencies (ICVA).

CAPACITY-BUILDING OF LOCAL PARTNERS

Current Policy and Practice

301. A series of policy documents and evaluations have highlighted the importance of capacity- building of local agencies. The CIDA Universalia report called upon UNHCR to engage in capacity-building. It was advocated by the CIS Conference (CISCONF/1196/6 4 July 1996) and the Advisory Group on Refugee Policies in Africa, convened in May 1996. PARinAC also made recommendations on strengthening of local agencies (recommendations 75, 86, 118, 119), and the theme has been echoed in the High Commissioner's speeches.

302. Although the importance of capacity-building is frequently emphasized, efforts to engage in it are limited. With the exception of Eastern and Central Europe, it has not explicitly formed a major part of any regional policy. It is usually undertaken as the result of personal initiative or becomes a theme when UNHCR wishes to phase out. In some cases UNHCR has been in a country for over 20 years without making any conscious effort to build up local structures.

303. Underlying the inconsistency of UNHCR efforts are the unresolved questions on the extent UNHCR should involve itself with capacity-building, or whether this should be left to development agencies. Capacity-building is a complex task requiring long-term engagement and special skills. Oxfam concentrates 80 per cent of its resources on capacity-building, but according to one of its Senior Officers, it has few successful models.

304. Despite the difficulties, four basic reasons have emerged to justify UNHCR involvement in capacity-building:

• the need to build up local agencies so they can act as viable partners;

• the need to increase the ability of local agencies to deal with refugee crises without external assistance;

• the need to build up local agencies to further reintegration; and

• the need to contribute towards the prevention of refugee crises by building up democratic institutions and those dedicated to the protection of human rights.

305. While there is some clarity on the goals of UNHCR involvement in capacity-building, there is no clear strategy regarding how it is to be undertaken and who should be targeted. In Eastern Europe there has been a tendency to focus on government institutions while in Central Europe local NGOs are targeted. This in part has been determined by the availability of partners and in part by individual preference.

306. With regard to the means to undertake capacity-building, at present the approach is often passive. Agencies are strengthened on the basis of the requests they submit (e.g., for vehicles, photocopiers, or computers). In addition, UNHCR does some training, usually in protection. A more proactive approach towards identifying and responding to the material and training needs of local agencies and government departments could yield more results.

Implementing Partner Training

307. There have been many calls for more training programmes, focusing on UNHCR programming methods, accounting systems and reporting requirements. While the coverage of protection training is extensive, this is less so the case with training in programming and financial management.

308. Where training is undertaken it is not always targeted to the right people or sufficiently repeated. Systematic training of those who actually write project descriptions and do the accounts is required. Where this has been done, compliance with sub-agreement clauses has increased significantly.

309. Training at the outset of a programme can pre-empt later problems from arising. To this end, the design of a crash training course for partners in emergencies has been proposed. Beyond training in UNHCR programming and accounting, training of local partners could also usefully include training in management and fund raising as well as, where appropriate, technical training.

310. The main constraint stated to UNHCR doing more training is lack of time and expertise. However, more resources may exist than are currently being tapped; efforts to fill two vacant Training Officer posts at Headquarters were long delayed. When not deployed in emergencies, the use of Emergency Officers (EPROs) and staff available under emergency standby arrangements for training of local partners (in technical areas, for example) could also be explored.

311. In a number of situations UNHCR has engaged specialised partners to train others. This was done with success with the European Committee on Refugees and Exiles (ECRE) and IBHI in Bosnia. In Armenia and Georgia this approach was taken a step further; local NGOs were established for the exclusive purpose of training and nurturing other local NGOs.

Mentor Arrangements

312. Given the limited expertise and human resources UNHCR can dedicate to capacity-building, seeking the active co-operation of other agencies is essential. In a number of situations international agencies have been encouraged to act as dedicated mentors to local agencies. In Bosnia a number of international agencies successfully nurtured local counterparts (e.g., Oxfam and BOFAM, the Marie Stopes Foundation and Stoppe Nada, IRC with an agency called CORRIDOR). Similar efforts have also been made in the Great Lakes region.

313. Such pairing arrangements thus far have occurred sporadically. Many think they should become the rule, and to achieve this have proposed the inclusion of a standard clause in sub-agreements, requiring international NGOs to engage in capacity-building of pre-identified local agencies and to hand over to them within a set time-frame.

314. There are number of constraints to this. In some circumstances local partners are hard to identify or may not be appropriate (see paragraph 98 in Part I). The extent to which international NGOs are oriented towards working through local structures also varies considerably. Not all international NGOs welcome the opportunity to phase out. In Sarajevo, for example, a reputable North American NGO engaged in community service activities has allegedly resisted handing over to local structures, despite this being written into the sub-agreement and local counterparts being available.

315. The constraints should not detract from the need for a more concerted effort to encourage international agencies to work through local partners and where possible to hand over to them. A discussion with international agencies on means of achieving this has already been initiated by the NGO Co-ordinator. The consistent application of the contribution of an agency to local structures as a selection criterion (see paragraphs 268-270 above) would also further this objective.

316. Another, related means of capacity-building, proposed by ICVA, could be through co-operation arrangements with other interested UN agencies. Capacity-building of local agencies is relevant to all international agencies, in particular in crisis-prone areas where local partners are lacking. UNHCR could seek the co-operation of other UN agencies and interested NGOs, in undertaking joint, multi-year projects aimed at identifying and strengthening a selected number of local agencies in crisis-prone areas.

INTERNAL CHANGES REQUIRED

Sub-agreements

317. The sub-agreement is the main document regulating UNHCR's relationship with implementing partners. Many believe the present format could be improved upon. Some suggestions put forward are:

• It should be made more like a commercial contract. It should contain performance targets and evaluation criteria (see paragraphs 291-293).

• The sub-agreement and its reporting requirements should emphasize project objectives as much as levels of expenditure.

• Different sub-agreement formats should be developed according to the size of a project. The present format could be substantially simplified for small projects (see paragraph 285). The possibility could also be examined of having different formats for local and international agencies.

• The budget structure should be made more congruous with those used by partners.

• Where relevant, clauses should be introduced on promoting protection concerns and/or undertaking monitoring (see paragraph 258).

318. A comprehensive review of the sub-agreement format could occur in the context of other changes to be made to programming instruments as a result of Delphi. This review could include looking at reporting requirements. The difficulties involved in gaining compliance with the reporting requirements spelled out in sub-agreements would indicate a need for change. Alternative proposals could be solicited from UNHCR's partners.

319. One constraint to making sub-agreements more objective-oriented and improving their value as contracts, is the lack of predictability of resources. UNHCR is in a weak position to demand adherence to performance targets unless it commits itself to providing, on a timely and reliable basis, the necessary resources. Where cuts are inevitable, they could be applied in a more differentiated fashion.

320. At present sudden across-the-board budget cuts, which appear arbitrary, undermine credibility and make planning difficult. They can also encourage more experienced NGOs to over-budget and in the longer term can prove expensive. In Belgium, for example, a request for a 25 per cent cut put in jeopardy an agreement negotiated by the Regional Office for a partner to take over all IC work with a commitment of three years' support from UNHCR.

321. UNHCR budget and accounting systems seldom interface well with the budget and accounting systems of governmental or non-governmental partners. The attempt to make them interface is labour intensive and costly. The UNHCR budget structure does not facilitate cost comparisons and the high level of detail it demands can encourage fudging. Accounting systems and budget structures could be reviewed on the basis of those used by selected governmental and non-governmental partners.

322. Efforts to simplify and make more transparent UNHCR programming procedures have been welcomed. The UNHCR partner handbook ("Partnership: A Programme Management Handbook for UNHCR's Partners") has proved popular and been praised.

323. Avoidance of the use of jargon in UNHCR programming language would save considerable time in training and explaining to partners. The Delphi tendency to replace old jargon with new ('LOI' becomes 'Envelope' instead of, for example, spending authority) should be avoided.

Role of Different Posts in the Selection Process

324. Given the crucial importance of implementing partner selection to the effectiveness of a programme and the large sums of money often involved, a broader consultation process would be appropriate. At present, Programme Officers play the primary role in implementing partner selection. They have, however, limited expertise and time to adequately undertake financial and technical assessment of partners.

325. Finance Officers: The role of Finance Officers in implementing partner selection and monitoring could be expanded. At present, their functions are focused on control of administrative expenditure, although such expenditure compared to project expenditure is minimal. The tasks they could undertake would include the following: ensuring potential partners meet financial selection criteria; monitoring of project expenditure; certification of SPMRs, and selective audit of partners.

326. An expanded role, embracing implementing partner monitoring, is foreseen for the new posts of Financial Management Advisors, which are to be allocated to Directors of Operations. These Officers could also be required to review and approve projects over a certain amount.

327. One of the main constraints to increasing the role of Finance Officers in implementing partner selection and monitoring is their scarcity. While the Lutheran World Federation, for example, aims at having an international Finance Officer in every operation, UNHCR with more complicated procedures and larger budgets, sometimes manages programmes of tens of millions of dollars with no international Finance Officer (in Guinea and Goma, for example). To overcome the lack of Finance Officers, in some operations such as Pakistan, JPOs and UNVs have also been used successfully to undertake financial monitoring of partners.

328. Technical experts: The role of technical experts could also be expanded in implementing partner selection and monitoring. Technical specialists could assess whether agencies have sufficient technical competence to undertake a given project. They could be required to sign off on any project, requiring technical expertise over a certain value. They could also be involved in monitoring and evaluation.

329. At present there are no clear criteria for postings or missions of technical experts. In programmes where they have been deployed, they are usually praised and their input is seen as crucial (see for example, the OECD DAC evaluation of the Rwanda emergency).

330. Technical expertise includes logistics. In 1996 UNHCR spent hundreds of millions of dollars on logistics and has built up considerable expertise. However, the experts do not always have a say in choosing partners in their sector. The envisaged post of Supply Chain Manager should be involved in selecting partners, monitoring and project evaluation in the logistics sector.

331. As with Finance Officers, technical specialists are often lacking. A September 1993 request to Headquarters from Mozambique for technical specialists was not met until early 1995. Greater use could be made of locally recruited technical experts or of experts seconded from government Ministries. This approach was used with success in Malawi.

332. The involvement of Finance Officers and technical specialists in the selection process would allow Programme Officers to focus on questions of programme design, programme objectives and overall cost effectiveness. Under the leadership of the Representative, Programme Officers could remain responsible for co-ordinating the selection process.

333. The attention paid by UNHCR Representatives and other senior UNHCR staff to implementing arrangements varies. In some cases, more emphasis is placed on the maintenance of harmonious relationships, in particular with governmental partners, than on the responsibility to establish the most effective implementing arrangements which the circumstances would allow for.

334. A reoccurring theme in Delphi discussions was the need to hold Representatives more accountable for the performance of partners selected. While the responsibility of the Representative for programme delivery and for the welfare of its beneficiaries needs continuous emphasis, it is hard to envisage concrete, workable mechanisms to enhance accountability in this respect.

Changes of Approach

335. Deficiencies in implementing arrangements can reflect an inability to develop a coherent programme strategy. A clear definition of the objectives of a programme and the needs of its beneficiaries is a prerequisite to an effective selection process.

336. In a finding highlighted by Delphi, in some cases UNHCR adopts a passive approach, waiting for partners to come to it with project proposals rather than seeking partners according to predefined needs (see paragraph 197 above). Where UNHCR arrives late or is thin on the ground this approach may be inevitable, but it limits UNHCR's ability to provide leadership and to give coherence to a programme.

337. In an ideal situation programme implementation consists of the following four stages:

• Determination of programme objectives and needs in co-operation with potential partners.

• Selection of partners in accordance with objectives and needs identified, and explicit, predefined criteria.

• Implementation of the project.

• Evaluation, in co-operation with the partner, of the project on the basis of objectives defined previously and predefined performance benchmarks.

A related approach was followed in Armenia. The UNHCR office, in co-operation with governmental and non-governmental partners, established a matrix of needs at the community level. Thereafter it attempted to find and match agencies to meet the gaps identified.

338. The ultimate purpose of the partnership between UNHCR and other agencies is to maximise resources dedicated to the assistance and protection of refugees and other beneficiaries. The approach to partners should thus be informed by two main concerns: the actual benefit which will result for beneficiaries from the partnership, and its value for money. All other issues - budget, reporting, audit, donor related - should be subservient to these concerns.