"The decision in October 1990 to formulate and implement an RBLAC Resources Mobilization Strategy was important in that it consolidated what until then had been ad hoc initiatives without clear goals by Country and for the Region as a whole. In addition, the Strategy established the mechanisms that allowed the RBLAC Directorate to assume full corporate responsibility and to exercise the necessary management discipline to move each and every one in the Region into achieving the Strategy's objectives. The Strategy was based on a reality for all to see, which followed the Fifth Cycle allocation of core resources. No matter what resources scenario the Governing Council was to come up with, Latin America would either be stagnant, or worse, decline, and would thus have to face the reality of IPFs too small to justify maintaining the network of RBLAC Country Offices as we know it. The possibility of having to consider closing Offices was one that RBLAC considered unacceptable, more so in light of its recent experiences, including the one of Central America, that had demonstrated the value of having a network of Country Offices in place, both in political as well as in developmental terms..."8

As a consequence, UNDP's role in Latin America has undergone a profound change over the last decade. The coincidence of the abrupt decline in IPF funds for the region and the dramatic shifts in socio-economic and governance situations throughout the region have significantly altered UNDP's relationships and operational circumstances. With these changes, some basic questions stand out. Are the co-financed programmes legitimate, i.e., compatible with UNDP's mandate and the objectives of multilateral cooperation for development generally? Do they bring added value to the host countries? Do the current UNDP infrastructure and procedures ensure efficient and effective delivery, and provide adequate guarantees against abuse?

There can be no doubt that the massive amounts of co-financing in the region have saved UNDP in Latin America from irrelevance. It has allowed UNDP to play a significant development role at a time when external cooperation-both technical and financial-seemed particularly important to the countries concerned. However, some precautions must be (and have generally been) taken to avoid abuses and, in particular, an excessive donor orientation of cost-shared programme action.

Regional Bureau for Latin America and the Caribbean Resource Mobilization Strategy for 1992-1996 (see footnote 8)

• Adopt a "Cooperation for Development" programme profile incorporating advocacy, programming initiatives (upstream), and support services.

• Spot stepping stones that lead to this profile.

• Maintain a network of country offices to promote an active dialogue on the Global Agenda-one of the major justifications for having a UN system presence at the country level.

• No country programme will be processed if it is not at least five times larger-including all additional resources, such as MSA, GEF, etc.-than the IPF.

• Limit as much as possible the approval of fully funded IPF projects.

• Use preparatory assistance as the operational vehicle to test the fund-raising potential of a "high priority" request.

• Demonstrate that UNDP can operate in a cost-efficient and effective manner.

• Seek out the country's development loans that are non-performing owing to national implementation limitations.

• Promote capacity-building through UNDP "implementation services" by allowing nationals to work with more apt rules and procedures and thus serve as incentives for advocating change in the rules and procedures of the country.

• Build on visibility and confidence gained from implementation services to move upstream to the identification and design of the country's loan portfolio and help to implement it.

• Promote "innovator-quality incubator" undertakings in experimental, upstream, participatory, sustainable human development approaches.

• Achieve a ratio in costings, expressed in terms of country office core budget: annualized 1992-1996 IPF, to be equal or less than UNDP's global core budget: IPF ratio.

• Create a cost-awareness at country-office level related to office budget requirement, handling more work with less staff (organization structure), staffing profile, "out-source" tasks.

• Decentralize authority on the management of country office non-core income.

• Establish effective information management system.

• National execution advantages for mobilizing non-core resources generated in-country:

- Enhances sense of ownership by the government;

- Speeds up the implementation process;

- Makes extensive use of expertise available in the country;

- Overcomes (temporary) lack of managerial capacity where otherwise qualified individuals in-country are available;

- Expatriates added for the high end of technical cooperation (short durations only).

Since 1990 for the RBLAC region as a whole, resource mobilization through co-financing has totalled eight times the Bureau's IPF for the fifth programming cycle with the remarkable mobilization of $1.9 billion for UNDP's co-financing operations while the IPF is $224 million. Adding the MSAs (not co-financing) of $381 million brings the total resource mobilization for the region to $2.3 billion. However, the variation among countries is considerable. By the fifth programming cycle (1992-1996), co-financing (excluding MSAs) in 13 of the 38 RBLAC countries exceeded country IPFs from 5 (Chile) to 50 (Argentina) times. In the remaining countries, co-financing was less than 5 times the IPFs or below the IPFs. Table 5 illustrates the variations in ratios of co-financing to IPF ranges.

The largest country participants in co-financing in the region are shown in table 6.

The distribution of the co-financing is extremely skewed. The country distribution of co-financing in the region shows that 65 per cent of total co-financing for the region is concentrated in four South American countries-Argentina, Brazil, Colombia and Peru-and one Central American country-Panama. Adding the six countries in the next group raises this share to about 83 per cent. The Central American countries, including Panama, represent only 14 per cent of total co-financing in the LAC region.

Sources of Co-financing.The ratio between IPF funds and co-financing and their distribution tell only part of the story. It would be erroneous to analyse co-financing by lumping together all the funds without considering in detail the sources from which they originated since the motives of donors, be they multilateral or bilateral, differ from one another as do the particular circumstances of each country that motivate the donors. Also, the rationales of the individual governments vary. In some countries the Governments prefer to enter into co-financing arrangements with UNDP rather than financing and executing their own projects out of their own budgets within the realm of their flawed public-sector organizations. The source of the non-core funds is thus significant, with important implications for UNDP's relationships with host countries, as illustrated in table 7.

Table 7 Summary of Main Sources and Co-financing Modalities for the Fifth Programming Cycle

Among the IFIs, the Inter-American Development Bank (IDB) has been particularly important in co-financing arrangements for the technical cooperation components of IDB loans. It is enlightening to dwell upon the significant role of IDB, not only because it is the regional financing institution but also because from its experience, one can draw inferences for other regions.

The relationship between UNDP and IDB started as an exchange of information about priority projects of special interest for both organizations. It was transformed into a close working relationship during the late 1970s and early 1980s. The Latin American debt crisis and changes in politico-economic circumstances make IDB an important lending institution in restructuring the region's economies, supporting macroeconomic adjustments, and reforming public-sector institutions and organizations.

At the same time that the Bank was shifting its emphasis, UNDP was facing a relative decline in its resources. In a way, the county offices' capacities for project execution activities were being underutilized. A symbiotic relationship began to emerge between the Bank and UNDP and resulted in the umbrella agreement.

The agreement is particularly interesting in pointing the way to a coordinated strategy between IDB and UNDP for the technical cooperation components of IDB loans. In one instance-Bolivia-a joint unit has been established to plan and manage the joint activities of the two institutions. For the regional LAC programme, which involved close cooperation on important advocacy action (environment, poverty alleviation, governance), periodic consultations between UNDP and IDB staff have helped in designing a joint strategy and optimizing opportunities for complementarity.

For most of the LAC countries, co-financing from government budgets (government cost-sharing) was the predominant modality during the fifth programming cycle. However, in four countries (Dominican Republic, El Salvador, Panama, and Paraguay) IFI loan funds constituted over half of the government cost-sharing resources.9 Also, in 1995 in Argentina, for example, total cost-sharing for the year is expected to amount to 99 per cent of the Argentina programme budget, or approximately $95 million as compared with $28 million in 1992. Of the 1995 cost-sharing, over $60 million is derived from IFI loans ($45 million from the IDB and $25 million from the World Bank.) Cost-sharing from government budgets has dropped substantially compared with previous years. In Peru, cost-sharing in 1995 amounted to over 95 per cent of the programme budget of $93 million. Sixty-seven per cent came from the Peruvian Government's own budget and 29 per cent from the IFI loans (24 per cent World Bank, 5 per cent IDB)10.

Bilateral co-financing has been far less significant in South American countries. It is important, however, in Central American countries wheretrust funds and bilateral cost-sharing have played a constructive role in support of peacekeeping and rehabilitation initiatives. In El Salvador, for example, the Trust Fund for Peacekeeping is financed mostly by Nordic countries, and in Nicaragua, Sweden is assisting, through a trust fund, in the settlement of property rights.

MSA arrangements are limited to Argentina, Bolivia, Ecuador, and Peru in South America and El Salvador and Nicaragua in Central America. In these countries they are used at present primarily for major equipment purchases and subcontracts. In Central America, an MSA is mostly used by the Japanese commodity assistance wherever the Japanese International Cooperation Agency (JICA) is not physically present. Contracts are carried out by UNOPS. MSAs do not, however, seem to represent a growth area.

UNDP-administered funds and GCCC play a minor role in both South and Central America. Parallel financing is also relatively unimportant or un-recorded in a number of countries. The most notable exception is Panama, where USAID has entered into a parallel financing agreement with UNDP within the "Modernización del Estado" (Government Modernization) project, which also covered the election process. There are hardly any regional projects in Central America financed by the cost-sharing modality. The only one is PRODERE, which has been completed. This project provided assistance to persons displaced by the civil strife in various countries of the region.

In summation, the predominant co-financing modality is cost-sharing with funds provided by IFIs through governments or by governments directly from their own budgets. In this situation, UNDP country offices have at times become heavily involved in project support services such as procurement, recruitment, and accounting.

International Financial Institutions as Sources of Co-financing. The proceeds of loans from the World Bank, the Inter-American Development Bank (IDB), and the Central American Bank for Economic Integration (CABEI) have been used for cost-sharing in the LAC region. The International Fund for Agricultural Development (IFAD) has also been present in cost-sharing in the region. Cost-sharing projects with the World Bank are governed by a standard annex while cost-sharing arrangements involving IDB loans are governed by an umbrella agreement negotiated between IDB and RBLAC. IDB's role in co-financing in RBLAC has been particularly important (see box on IDB participation).

The World Bank's objectives coincide with those of UNDP in areas such as public-sector reform, stabilization, structural adjustment, and poverty alleviation. The World Bank in its loan agreements deals directly with the Governments, which may then choose to channel some of the loan funds for technical cooperation to projects using the cost-sharing modality with UNDP. The Governments may choose UNDP provided there is compliance with the "no objection" clause of the World Bank. As pointed out by a World Bank official knowledgeable about co-financing arrangements, the World Bank views UNDP's role as particularly important in project implementation and related capacity-building. Upstream work of a substantive nature is considered a strength of the IFIs with considerable expertise of their own to draw on; thus, they are in less need of UNDP's participation in that work. There are indications that UNDP programme action-whether with core or non-core resources-is increasingly used for preparatory assistance and to build up the capacity of the countries to formulate their own borrowing policies, priorities, and loan proposals. Many recipient countries of the region attach importance also to UNDP's objectivity and expertise in the broad area of governance and institutional reform-an area in which relations between IFIs (particularly IDB) and UNDP have, in fact, been evolving towards an increasingly close partnership.

However, the relations between these IFIs and UNDP have varied considerably from country to country. In Peru, UNDP helped develop a loan- project pipeline when the IFIs were not present in the country. This situation continues with UNDP involvement in the formulation stages of IFI loan projects. In other countries, UNDP (as well as the national coordination agency responsible for technical cooperation) is consulted only when the loans have already been formulated by the IFI and the national agency responsible for external borrowing. In general, where UNDP is to be involved in the technical cooperation components of a loan project, its participation in project formulation is considered desirable. In other situations, UNDP has used its core funds or cost-sharing to help in the preparation of loan projects for national, state, and municipal governments.

The recent assessment of UNDP-IDB working relationships now being completed provides useful insights on the strengths and weaknesses of UNDP-IDB collaboration. Its analyses and lessons should be drawn on to guide UNDP relationships with other development banks in Latin America and other regions in addition to those with the IDB.11

As indicated in tables 6 and 7, the Governments, including IFI loans, are the most important sources of funds for co-financing arrangements via cost-sharing modalities. What have been the factors that have favoured the use of this modality on such a large scale?

Several factors were at work in the LAC region during the late 1980s and early 1990s, i.e., towards the end of the fourth programming cycle and in the beginning of the fifth one, that favoured co-financing arrangements. These played a vital role in the profound transformation that took place in practically all the LAC countries. That transformation, in turn, opened wider the window of opportunity for UNDP. These factors were:

Political Factors. Military regimes have been replaced by civilian governments, e.g., Argentina, Brazil, Panama or Paraguay, or decades-long civil strife came to an end, as it did in El Salvador, Nicaragua or Peru, for example. Authoritarian regimes were replaced by pluralistic democracies, however fledgling they may be, which made imperative the identification of development priorities with the participation of vocal civil society. Long-neglected goals of public-sector reform, capacity-building, poverty alleviation and human resource development occupied a place of the utmost importance for the new Governments. It was essential not only to revitalize and realign the economy but also to redesign systems of governance and minimize the impact of the social cost of the structural adjustments which had serious political implications. UNDP, as a neutral agency, was involved in identifying such goals with precision, i.e., in the upstream dialogue, and converting them subsequently into projects.

Economic Factors. Political changes also led to economic changes. The economic sectors needed reconversion and adaptation, but more importantly, the whole economic structure needed realignment through privatization and debt conversion. In many countries, the role of IFIs was reduced to a bare minimum because of the payment arrears or because the decision-makers decided not to deal with them since the IFIs' strict rules of the game did not suit their purposes. The countries needed urgent assistance at all levels, and they could turn to UNDP. The involvement of IFIs had to wait until the arrears were cleared.

Weak Institutions. The new Governments took over institutions that were weak for a variety of reasons. In Nicaragua and Panama, for example, there was a diaspora of intellectuals, highly technical civil servants since they could not, or did not want to, collaborate with the dictatorial and/or corrupt regimes. The outmoded institutions had not been overhauled since such reforms did not suit the purpose of the previous regimes. In other words, the Governments began to function with a flawed institutional setup. The expectations of the civil society, on the other hand, had to be fulfilled as soon as possible. UNDP was the only international entity in situ that could cooperate with the new democratic Governments in order to ensure that formulated goals could be met.

Not all of these factors were present in all the countries all of the time; they played different roles but opened windows of opportunity for UNDP in most countries of the region. In conjunction with these fundamental transformations in country circumstances, several specific concerns were at work favouring co-financing arrangements. These vary from country to country and can be summarized as follows:

• the country's bureaucratic system and laws and regulations are so cumbersome that the speed with which certain projects need to be executed can be better achieved through UNDP;

• the public sectors do not have adequate human resources required for effective project conceptualization, formulation, and execution since capacity for such activities has yet to be built;

• the Governments view UNDP as a neutral, transparent, and corruption-free organization. Decision-makers may prefer UNDP execution to their own public-sector execution; and

• where countries have suffered from political strife and where UNDP has appeared to have been instrumental in establishing peace, certain specific projects are entrusted to UNDP, which enjoys the confidence of diverse political factions.

These reasons may also be valid in the case of co-financiers other than national governments.

Not all bilateral donors are equally active in all the countries of the LAC region. Some bilateral donors do not enter into cost-sharing arrangements with UNDP. They maintain large missions in the countries to manage their own projects. Examples are JICA in Panama and the United States Agency for International Development (USAID) in Central America and Bolivia. By and large, however, there are five reasons why bilateral donors enter into co-financing arrangements with UNDP:

• a very small staff nucleus in the country, hence inability to manage and monitor projects;

• a conglomeration of small projects ($ 100,000 or less) that may lead to inefficiencies if managed by diplomatic missions in the country or from the headquarters in donor countries;

• the view of several bilateral donors that financing a project or programme managed by a neutral entity (UNDP) is more efficient;

• the language barrier that arises from the difficulty of recruiting staff fluent in Spanish or Portuguese (the case notably for Nordic bilateral donors); and

• a preference for participating in projects managed by UNDP that are politically sensitive but that are recognized as critically important to a country's evolution to peacekeeping, democratic systems and economic reform.

In this environment, UNDP had the opportunity to provide distinctive services at critical moments:

• RBLAC countries' perspective. From the perspective of the Latin American countries, the value added by co-financing with UNDP in cost-sharing modalities for country projects resides primarily in a more efficient utilization of existing resources and their management in line with development objectives, whether the resources come from the Government's own budget or are derived from IFI lending. Government or the IFI-derived cost-sharing does not, by itself, bring additional resources into the country-except to the extent that UNDP involvement may be a condition for, or facilitate the negotiation of, IFI loans.

• IFI perspective. From the perspective of IFIs, co-financing arrangements with UNDP help improve loan performance, especially in areas where disbursement rates have traditionally been low, e.g., in the social sector, and in major institutional reform programmes. Better disbursement rates are, of course, also of interest to host-country borrowers, who are often paying substantial commitment fees for non-performing IFI loans.

• UNDP perspective. Finally, from the perspective of UNDP, co-financing represents the only way to have an impact in countries with core IPF resources as modest as those assigned to the region. In that context, it opens the door for UNDP-through a continuing policy dialogue with the host countries-to advocate policy priorities such as the concept of sustainable human development. Fees charged in connection with co-financing help to staff viable country offices, which in turn are a condition for successful programme implementation and thus represent one of the main comparative advantages of UNDP.

The emergence of massive cost-sharing in the LAC region brought with it a fundamental shift in the nature of technical cooperation. The bulk of today's programme action, especially where government or IFI cost-sharing is involved, corresponds to a concept of cooperation that often focuses on such sensitive areas as governance and national policy. These call for institutional capacity-building and investment in orders of magnitude far in excess of the modest allocation of core UNDP resources. For example, in Panama, projects to modernize the State utilize 38 per cent of IPF funds but 54 per cent of the total IPF and cost-shared funds.

The opening of windows of opportunity per se does not imply that the opportunities will be taken advantage of automatically. In the particular context of Latin America, both the regional bureau and the country offices took specific steps to exploit such opportunities. The country offices, in their advocacy role, gained the confidence of the country and the respect of the decision- makers. Several country experiences attest to this fact. In El Salvador, for example, both the Government and the rebel forces requested UNDP's assistance in the post-civil war period (see box on the peace agreement and UNDP). After the fall of the dictatorship in Panama, UNDP was at the forefront, helping to build a national consensus.

Second, while many multilateral and bilateral donors stayed on the sidelines during the turbulent years, UNDP country offices were the only entities with institutional memory ready to take the initiative and capable of assisting the Governments in shaping development policies and strategies. Equally important, they were capable of formulating programmes, designing projects and, with limited core funds, initiating preparatory assistance for larger programmes/projects that were suitable for subsequent donor financing. They were thereby able to strengthen UNDP's participation in national development. The experience in Peru is very illuminating in this respect. UNDP was the only multilateral agency present during the country's profound economic and political crisis. It assisted the Government in developing a portfolio of bankable projects against the day, which came in 1991, when the IFIs would resume their lending operations in the country.

Third, the country offices geared themselves to provide support services to implement many government projects and programmes without delay. Thus, to the extent that the demand for UNDP assistance was revealed, UNDP supplied, in a timely fashion, a set of "products" without which the new Governments could not have functioned as well as they did. Many resident representatives played their advocacy role with a great deal of common sense and perception, and demonstrated the necessary political savvy and managerial capability.

Given the large quantity of additional resources that will be needed to implement the peace plan, both sides call upon the international community for assistanceA National Reconstruction Fund will be created with the support of UNDPThe role of UNDP will be to assist in the mobilization of external funds, preparation of projects and programmes that will have the support of the donor community, expedite the process of obtaining technical assistance from multilateral and bilateral donors, and to collaborate with the Government in harmonizing the reconstruction plan with the activities of the NGOs dedicated to local and regional development. (Translated from: Peace Agreement, signed in Chapultepec, Mexico on 27 April, 1991 by the Government and FLMN, in Naciones Unidas, Acuerdos de El Salvador: En El Camino de la Paz, San Salvador, 1993, p. 89.)

In summary, the primary factors which seem to have contributed to successful co-financing arrangements in the Latin America region are:

• the political support of the host country;

• the support of IFIs, to the extent that cost-sharing is to be derived from their loans, and keeping in mind that their basic interest is to improve loan performance in areas (social sector and governance) where IFI performance has traditionally been poor;

• the absorptive capacity for external cooperation: in Latin America, delivery rates have always been very high, no doubt owing to the nature of the cooperation and the role played by the country office;

• the capacity of the country office to manage and monitor complex programmes and projects;

• UNDP staffing policy; and
• the legal and administrative culture of the host country.

Furthermore, in Latin America, there has traditionally been a high degree of empathy between the UN system, UNDP, its Regional Bureau and the country offices on the one hand and the countries of the region on the other.

In many country programmes, cost-sharing has added substance and impact to social-sector and environmental management projects responsive to SHD. Gender equality and the guarantee of human rights, employment generation, health and education also play a key role in this connection. The emphasis is on upstream measures, i.e., on institutional capacity-building and organizational reforms by which the country can design and implement its own policies to combat poverty, preserve the environment, and exploit its resources in a sustainable manner. Employment generation, education and health play a key role in this connection, as does good governance in which decentralization and devolution of power to civil society coincide with the emergence of new roles for the State. Hence, public-sector reform and capacity-building in the public sector are top priorities. In fact, these correspond to UNDP's programme stand since UNDP is a key advocate of sustainable human development which presupposes capacity-building in government and in civil society, i.e., creating or reforming the enabling environment. The Governments' priorities, in more cases than not, did and do coincide with UNDP's stand.12

It is important to note that these priorities have overlapped with those of IFIs as well, which, after a hiatus, renewed their activities in Latin America. The Governments had to accept certain conditionalities for new loans and assistance after the arrears had been cleared. These included public-sector reform, liberalization of the eco-nomies, poverty alleviation, and similar social programmes, precisely the ones that UNDP was advocating.

It was not difficult to see, then, why the IFIs were quite willing and perhaps eager to co-finance several projects with UNDP. Similar arguments of overlapping priorities also apply to bilateral donors where there is a convergence of development priorities. Building a consensus among governments and donors on these priorities and on specific programmes to address them is an appropriate role for the UNDP country office.

The regional programme has two important functions:

• advocacy, where it has played a precursor role both in the area of poverty alleviation and in gaining the support of the region for the concept of sustainable development and the preservation of the cultural patrimony; and

• managing complex multi-country projects in critical areas such as transport, the introduction of competitive industrial technology, economic policy and integration.

In both types of projects, the Regional Bureau has been able to mobilize cost-sharing for the fifth programming cycle amounting to approximately $8.1 million. For example, the IPF input for the planning and engineering studies for the environmentally sustainable development of the Parana-Paraguay-Plate River waterway-an initiative of critical importance for the five countries of that river basin-resulted in substantial cost-sharing with the IDB. In other cases, e.g., civil aviation projects, cost-sharing has come from governments, including the Government of the United States. Spain and Sweden provided cost-sharing for projects in the area of governance. Co-financing or parallel financing from the IDB was obtained for important advocacy efforts at policy levels, in particular in the areas of poverty alleviation, environmental management and governance. It is difficult for the IDB and the World Bank to make multi-country loans. UNDP regional projects, supported by cost-sharing derived from IFI loans to the individual participating countries, are thus seen as a possible approach to the IFI financing of regional or multi-country development.

The LAC experience with co-financing on a large scale in most of the LAC countries has raised a number of important management issues that need to be carefully considered as these arrangements are continued.

Importance of Core IPF Inputs. Although in quantitative terms, core UNDP resources represent only a few percentage points of the overall programme budgets, they have played an essential role even in countries with high cost-sharing ratios, where many important projects are funded entirely from non-core resources ("one hundred per cent cost-sharing"). On the one hand, core IPF contributions lend credibility to UNDP participation in national development. On the other hand, core IPF funds-even in very modest quantities-have been used for preparatory activities or as seed or front money in projects with substantial amounts of cost-sharing. This capacity is important in maintaining UNDP's independence and integrity as well as its role in substantive matters beyond project support services which are covered by fees from co-financing arrangements.

The use of UNDP core funding for preparatory activities deserves particular attention. The 1990 UNDP/IDB umbrella agreement recognizes the importance of core IPF "seed" inputs and specifies that they may be considered as advances reimbursable out of the proceeds of eventual IDB loans. Some World Bank staff have noted that, whether or not formally recognized as a reimbursable advance, UNDP core funding for project preparation was a more agile modality than the World Bank project preparation facility (PPF); in several instances, it has been used to help the borrowing country prepare loan projects. In these instances, the technical cooperation component of the resulting loans will, in principle, be handled by UNDP under cost-sharing arrangements. Examples in which small amounts of core IPF funding served to prepare the ground for substantial cost-sharing include the privatization programme in Peru and large social-sector projects in the other countries.

Autonomy and Control: Preventing Abuses. Two basic concerns have been voiced with regard to massive cost-sharing. One is that the availability of cost-sharing may lead UNDP to approve projects to which it can make no substantive contribution. It is thus feared that UNDP may be used for purely administrative functions (i.e., a simple crown agent or fiduciary role), as a hedge against currency fluctuations, or as a convenient device to bypass national tax legislation or generally applicable limitations on public service staffing and salaries. A related concern is that cost-sharing, especially government cost-sharing combined with national execution and the use of national consultants, may serve to finance ordinary government functions, generating durable dependencies and a two-track civil service-neither of which would be compatible with good governance and the concept of sustainability.

In fact, it was found that both governments and UNDP country offices have become more aware of these problems. Occasionally a project may have been approved primarily to establish a relationship of trust with the host government, and conjunctural motivations may at times also have been present in cost-sharing arrangements.

However, in the four South American countries visited, for example, the UNDP country offices have rejected important cost-sharing project

proposals when they were not satisfied about their soundness in terms of substantive content, modalities of execution or sustainability or if they felt that UNDP would have no substantive role and no means of exercising actual management control. In most instances, the project design was ultimately modified in conformity with UNDP suggestions-often based on the results of a preparatory project or on the advice of external experts retained for that purpose. In a few cases, projects or project components in which UNDP would have played no substantive role were converted into trust fund or MSA arrangements with specialized agencies or UNOPS.

Where cost-sharing comes from the government budget, the UNDP country office generally participates in the project identification and formulation, and it is thus rare that proposals have to be rejected. Where cost-sharing is derived from IFI loans, on the other hand, the country office and its national counterparts charged with the coordination of technical cooperation are rarely consulted at that stage. An exception is Peru, where for historical reasons, UNDP is involved by the Government in the formulation of many IFI loan projects. It is noted that with regard to the LAC regional programme, the 1990 UNDP/IDB umbrella agreement provides for consultations between RBLAC and IDB headquarters at the prospective stage, when priorities and projects are identified and formulated. The UNDP field staff feel that this should also be the case at the country level, for IDB and World Bank lending as well. To that end, an informal but continuing UNDP/IFI dialogue at the country level would probably be sufficient but should also include, to the extent possible, the national counterpart institutions charged with the coordination of external financial and technical cooperation, respectively. In special cases-as for the IDB/UNDP cooperation in Bolivia-a joint coordination unit may be necessary.

The danger of generating durable dependencies and a two-track civil service by using cost-shared UNDP projects to hire large numbers of staff for government departments in the four South American countries visited appears to be under control. While the use of national consultants under UNDP contracts can be an important temporary expedient in the context of a deliberate institutional reform or capacity-building strategy, the abuses that may have occurred in the past are today avoided by prudent programme design and management at the country level, by statutory limitations on international contracting for government jobs, as prescribed by current Argentine budget legislation or-as in Peru-by entrusting supervisory or control functions to an interministerial body in which the UNDP country office has an active voice.

Co-financing and Accountability. Two aspects of accountability should be viewed separately: financial and substantive. There are no indications that financial accountability is compromised in the projects implemented through the NEX modality, and almost the totality of them are co-financed in the LAC region although audits may be necessary to confirm this conclusion.13 The issue of substantive accountability is somewhat complex, simply because the country offices are overburdened with large numbers of projects, and programme officers usually carry the responsibility of both financial and substantive accountability. Financial accountability always tends to supersede substantive accountability for obvious reasons. So far there is no evidence that substantive accountability is compromised in the LAC region either. One potential danger is that, especially in projects where cost-sharing is with the Government, monitoring and evaluation are viewed to be somewhat superfluous. So far, evaluations prescribed by the project documents have taken place. However, should the Governments insist on their view, there may be a danger of compromising substantive accountability in the future.

Cost-sharing, National Execution and the Programme Approach. National execution (NEX) is now the norm of project execution following General Assembly Resolution 41/199 and Governing Council decision 93/25. In other words, during the fifth programming cycle, most of the projects-co-financed or not-have been or are being implemented through this modality, unless very compelling reasons to leave the implementation to one of the specialized agencies prevailed. Although the norm, the question of whether or not NEX is conducive to co-financing still persists.

The LAC experience indirectly indicates that NEX is an inducement for cost-sharing. When a project is cost-shared with the Government's own funds, there is a tacit understanding that the modality will be NEX. If the government funds for co-financed projects originate from IFI loans/grants, i.e., governments in fact are left free to choose the implementation modality (subject to the "no objection clause" of the IFIs), the tendency is always towards the utilization of the NEX modality. The bilateral donors usually accept the implementation modality suggested by UNDP, which is NEX.

In the Latin American countries reviewed, cost-sharing is closely associated with national execution and the use of national consultants and with the priority attributed to capacity-building. The link with national execution is not exclusive, however; in the Colombia programme, for instance, several important projects are executed by UN system agencies. Nor does cost-sharing impede cooperation agreements with the agencies in the implementation of nationally executed projects. In some instances, cost-sharing is channelled directly to trust funds established with them.

Similarly, the programme approach provides a useful matrix for the formulation, and especially also the monitoring and impact evaluation, of large cost-sharing projects or clusters of projects. It is, in fact, considered that UNDP action, whether or not it involves cost-sharing, must relate to a broader universe of needs and resources-both internal and external-applied to the development process, either countrywide or broken down by theme or area.

Cost-sharing Fees, Disbursements and Administrative Problems. Cost-sharing involves a negotiated fee ranging from 1 to 3.5 per cent, charged to administrative and operational services (AOS); 80 per cent of this fee remains with the country office while 20 per cent is channelled to UNDP Headquarters. These charges have occasionally been questioned both by government and media in the host country and by IFI staff. As long as it is clear that cost-sharing is not simply a convenient device for bypassing national legislation or a hedge against currency fluctuations (which today play a minor role in the region) and that UNDP is in fact making a substantive contribution in terms of management, monitoring and project support, it has been relatively easy to demonstrate that the cost-sharing charges are justified and correspond to services actually rendered by UNDP, especially at the country office level.

No major difficulties are reported with regard to cost-sharing disbursements. Delays have at times occurred in the processing of withdrawal authorizations by IFIs, with the issuance of "no objection" letters, or where disbursements have to come from ministries, States or municipalities with solvency problems. In all these instances, prudent management, and in particular financial monitoring in real time, must ensure that no commitments or expenditures are undertaken unless the corresponding cost-sharing has been paid. The alternatives are for UNDP to claim the accrued interest on cost-sharing (which otherwise would flow back into the overall programme budget) or then to charge a fee line item by line item. In appropriate cases, core IPF resources allocated to the particular project may of course be used as front money pending the payment of cost-sharing, but there, also, prudent management and monitoring are of the essence.

In many instances, government cost-sharing may be paid in local currency. However, care must be taken by country office management and central UNDP controls to ensure that the programme as a whole can absorb these local currency payments without undue delay, failing which UNDP would be exposed to the risk of currency fluctuations and limits on convertibility.

Some problems of an administrative nature are reported, especially with IDB-derived cost-sharing or MSAs. They concern, for instance, the cumulation of auditing and reporting requirements and the different exchange rates applied by IDB and UNDP (daily vs. standard monthly rates). As regards the World Bank, most administrative problems are resolved by Standard Annex II appended to all loans involving UNDP cost-sharing.

The UNDP Country Offices. Accomplishments in resource mobilization through co-financing have been, to a large extent, the result of the initiatives of those UNDP country offices that were able to move flexibly and creatively and establish a relationship of trust with the host country. These initiatives were spurred on by the RBLAC's management strategy that has been outlined at the beginning of this section. Effective policy dialogues between the Government and UNDP have been, for the most part, a consequence of the quality of UNDP field representation-dialogues upon which the partnership concept, the effectiveness, and also the legitimacy of UNDP action are predicated. It has been incumbent upon the country offices to identify and negotiate opportunities for cost-sharing, to avoid possible abuses, and to manage and monitor portfolios of large and complex projects with a considerable degree of autonomy. These are functions that cannot be easily legislated or regulated from headquarters.

In the countries reviewed, country offices have had to operate with very small, regular manning tables, relying to a large extent on staff financed by cost-sharing fees or other add-on funds. It is in fact doubtful that without this additional staff, the country offices would be able to fill the role they are expected to play. As matters stand, it was found that despite heavy work loads, they have done remarkably well in programme development and project formulation: the programme approach has no doubt contributed to the quality and clarity of the project documents. The country offices have also been remarkably successful in designing and operating their own financial monitoring systems, using state-of-the-art informatics technology which allows real-time control at both the project and programme levels. In the opinion of governments but also of IFIs, the quality of UNDP monitoring is in fact one of the important values added by cost-sharing arrangements. Continued substantive support and impact evaluation, on the other hand, have stretched the staff resources of the country offices to the limit. Increasingly, country offices have had to resort to external experts, retained on an ad hoc or intermittent basis and funded from cost-sharing fees and other add-on funds, to follow and evaluate some of the more complex projects in their portfolio.

There is an important lesson in this for UNDP personnel management and in particular for the career development of its professional staff. In order to do justice to a mature, partnership-based model of cooperation involving, at times, very large amounts of financing, resident representatives and their senior assistants must not only be good managers, but also they must be able and prepared to empathize with the country in which they serve and entertain close, policy-level relations with their national counterparts. This may-and has in fact-pose some problems insituations of political discontinuity. In such instances, much depends on the personal qualities and sensitivity of the resident representative and, ultimately, on staff deployment decisions taken at UNDP headquarters.

It is difficult to foresee whether co-financing can continue as a permanent funding arrangement for UNDP in the LAC region. It is likely that it will continue in the region during the next programming period with variations among the countries. There are several indications as to why this may be the case:

• Capacity-building in the public sector is not yet nearing its completion. Formulation and implementation of a number of projects and many management functions will have to be left to the UNDP country office.

• Two important IFIs-IDB and CABEI-are interested in tackling capacity-building and social projects. Forthcoming loans from these institutions will likely include such components in the agreements. Neither institution sees itself as an expert in these fields. Their past positive experience in working with UNDPin these areas and the in-house capacity of the country offices will not only facilitate co-financing but will also induce it, as representatives of both banks in Central America noted.

• Many bilateral donors are not interested in expanding their staff in LAC, especially in Central America. Hence, they are likely to co-finance projects to a limited extent where special situations call for close collaboration through co-financing.

There are also factors that work on the negative side for the continuation of co-financing:

• As conditions become normalized and IFI loans move with the expected speed, the urgency with which UNDP's help has been sought in the past will not exist and may no longer be a determining factor.

• Assisting with the rapid transition from a dictatorship/command economy to democratic regimes/a market economy is a priority for many bilateral donors, which may wish to manage their assistance directly. The UNDP country office had remained in these countries as the only viable international organization. After the normalization of conditions, priorities will change and a hefty reduction in the flow of funds may be observable. Cases in point are the large Japanese cost-sharing funds for privatization in Peru or trust funds to implement the peace process requirements in El Salvador.

• Most important, to the extent that the coun-tries reform their public sectors and build the human resource capacity to manage them, there will be a diminishing need for UNDP country office assistance in formulating and managing technical cooperation programmes and projects. UNDP is heavily involved in both areas. If it succeeds, if capacities are built and the public sector is reformed, cost-sharing should diminish. If not and cost-sharing continues, then the implication would be that such capacity-building programmes have not been successful and will not be sustainable.

• Finally, questions must be raised with respect to MSAs. Some bilateral donors expressed their dissatisfaction with the management of these agreements that are handled mostly or exclusively by UNOPS. Should this trend continue, some bilateral donors, especially Japan, are likely to search for other modalities.

The question may be asked whether, given the emphasis on institutional capacity-building and governance in technical cooperation, the day will come when UNDP programme action and related cost-sharing will become redundant. While institutional self-sufficiency is the ultimate objective, one can visualize that there will always remain a need for an institution capable of objectivity, transparency, and global outreach to manage complex development programmes. UNDP, as an international organization, combines these characteristics. Hence, one can say that during the next programming period, co-financing is likely to continue. A dramatic upward or downward change should not be expected although changing circumstances in the major co-financing countries would have a substantial effect on UNDP's operations. This estimate is based on two assumptions: one, that UNDP cooperation will continue to focus on governance, institutional capacity-building and management; and two, that UNDP will maintain a country presence and continue to assign some core funds, even in modest quantities, to play an innovative, catalytic and seed role.

At this point in time, the same statement cannot be made with the same degree of assertiveness for the subsequent programming period. The continuation of co-financing will depend upon the extent to which UNDP, the IFIs and bilateral donors share common policy perspectives, adopt the same stand on SHD and poverty alleviation, and are able to influence recipient government policies and programming accordingly. Where this is the case, UNDP will likely continue to have opportunities for managing co-financed programmes serving SHD and poverty-alleviation goals.