The Multilateral Donor Non-System: Towards Accountability and Efficient Role Assignment

Global governance refers to several pillars; one important pillar is the multilateral aid architecture. Its reform can be discussed under the perspectives of representativeness, inclusiveness, and efficiency (of aid delivery). A prerequisite for efficient aid delivery is to map the rising complexity of multilateral development finance, to help identify areas for consolidation, address fragmentation and poor co-ordination at country level, and help identify comparative advantages for an institutional role assignment among multilateral agencies. After doing just that, the paper explores why the multilateral donors have proliferated and provides broad recommendations for a more efficient and accountable multilateral donor system. --


Introduction
The rise of emerging countries with large populations and energy resources, arguably the dominant economic impulse of the 21 st century, and the current grave financial and economic crisis generate a "unique opportunity for creative diplomacy" (Kissinger) for a new world order. Global governance refers to several pillars; one important pillar is the multilateral aid architecture to describe institutions and instruments to support global development and the supply of global public goods. The rise of emerging powers as new official donors, the creation of vertical funds to finance global health and other global commons, the growing role of non-governmental organisations and the increased presence of private philantropy have triggered a growing debate on the need to reform the multilateral aid architecture. Developing country recipients complain about the administrative burdens imposed on them by donor missions, evaluation bureaucracies, policy dialogues and other strings attached to aid money.
Global governance reform can be discussed under the perspectives representativeness, inclusiveness, and efficiency (of aid delivery): • The debate on the reform of the Bretton Woods institutions has been defined by the need to restore representativeness to the system to reflect the shift in economic and political clout toward the emerging countries; this perspective is familiar to a broad audience by now and extensively debated in the literature already (see, e.g., Boughton and Bradford 2007). IMF quota reform is now underway. • The apparent shift in global governance from the G8 to the G20 in the wake of the global financial crisis 2007-09 has alerted some observers to the fact that global governance fails to be broadly inclusive even after that shift. The challenge is to find ways to have small countries participate in global governance, either through double-majority (budget and number of countries) voting principles, through delegated voting mechanisms or ad hoc coalitions (Narlikar and Tussie 2004).
• The efficiency of aid delivery has been emphasised in the Paris Declaration 1 . The Paris Declaration notes that excessive fragmentation of aid at the global, country or sector level impairs aid effectiveness. It called for increased donor complementarity to reduce transaction costs. Further, the Paris Declaration wants to promote a model of partnership that improves transparency and accountability on the use of development resources. It recognises that for aid to become truly effective, stronger and more balanced, accountability mechanisms are required at different levels.
The Accra Agenda for Action (AAA) commits donors and developing countries to "complete good practice principles on country-led division of labour" and to "start dialogue on international division of labour across countries by June 2009". This puts division of labour firmly on the agenda for multilateral donors just as much as for bilateral donors. As for multilateral aid, this efficiency aspect is little discussed so far, with the possible exception of streamlining UN aid under "One UN" (Chataîgnier 2008;Vatterodt 2008).
The 2009 OECD/DAC Report on Division of Labour (OECD 2009) examines the concept of aid fragmentation across countries and what has happened since the adoption of the Paris Declaration. It also proposes measures for concentration and fragmentation and options for tackling excessive fragmentation. The policy principle is that where a donor/ partner aid relation is particularly neither significant from the donor's point of view, nor from the recipient's point of view, there is an opportunity to achieve some rationalisation. In a "narrow" definition of concentration measures, the OECD score measures whether a donor provides a higher percentage of a recipient's aid than it provides of total global aid (the donor is "concentrated"), and whether the donor is also among the largest donors that together account for at least 90% of all aid to the recipient (the donor is an "important" donor). On average, one third of bilateral donors' relations fell in this group in 2008. For multilateral agencies, the concentration ratio was a little lower _________________________ 1 The Paris Declaration, endorsed on 2 nd March 2005, lays down a roadmap to improve the quality of aid and its impact n development. 56 partnership commitments are organised around five key principles: ownership, alignment, harmonisation, managing for results, and mutual accountability. (http://www.oecd.org/document/18/0,2340,en_2649_3236398_35401554_1_1_1_1,00.htm l) www.economics-ejournal.org on average (26%), mainly because of the global mandate of many small agencies; the trend was negative as the OECD's narrow concentration measure for multilaterals fell from 29% in 2004.
Given the limited progress achieved toward less multilateral donor fragmentation since 2005, this paper is more radical: In order to promote accountability, it is proposed to assigning UN Millennium Development Goals (MDGs) (goals and/or targets) to multilateral agencies together with a specialisation among them along the lines of the Tinbergen rule.
In line with the objectives of the Paris Declaration this paper is concerned with the efficiency of aid delivery by multilateral donors; in other words, it works toward reducing redundancies, mission creep and overlap so beneficial for economists' and ex-politians employment but so burdensome to poor countries (and donor budgets). However, this paper is far from producing operational instruments that would foster devision of labour and accountability among the multilateral donor organisations. To be sure, donors cannot be made fully accountable for the attainment of a specific MDG, because it is usually influenced by many factors that cannot be controlled by them. Nonetheless, the MDGs have the advantage to break down into targets so that may actually facilitate traceability; moreover, they are timebound which opens the opportunity to try different organisational responsabilities over long time spans.
The next section will map the rising complexity of the donor landscape. Section 3 will then look for positive and normative explanations of the rise in multilateral complexity. Section 4 will briefly document the cost of complexity to donor budgets and to recipients. Section 5 will try to identify avenues toward efficient role assignment.

Mapping the Multilateral Donor Non-System
The international development-finance system has become highly complex. New actors, both public and private, have emerged as important sources of finance. Traditional donors have begun using new financing instruments to deliver their aid. The goals of development assistance-already numerous-have broadened to include global and regional public goods. One would like to think that the www.economics-ejournal.org international aid architecture is an orderly process guided by simple principles, but the trends clearly show that we have a non-system 2 . Unlike some of its elements (such as the Bretton Woods sister organisations), this non-system does not result from coherent design, but is a child of spontaneous disorder. In recent years, the multilateral aid system has been growing in complexity. Too many multilateral organisations, with overlapping mandates, complex funding arrangements and conflicting requirements for accounting and reporting seem at odds with the aid effectiveness agenda.
The multilateral donor non-system needs serious mapping, a time-consuming exercise that is only now being started at the OECD/DAC 3 . What is the rationale of mapping? Such mapping identifies overlaps-leading to reduction of multilateral remit or proposals for consolidation; rivalries -leading to clarification of roles; and absences of co-ordination-leading to the design and implementation of co-ordinating structure. The mapping of the multilateral landscape is required to help identify areas for consolidation, address fragmentation and poor co-ordination at country level, and help identify comparative advantages for institutional role assignments among multilateral agencies. These are preconditions for the multilateral system to deliver aid effectively, with benefits for donors and partner countries alike.
The DAC's Creditor Reporting System (CRS-database of aid activities) allows for recording the 'channel of delivery' in order to collect information on aid routed through the multilateral system. It thus provides a natural starting point for mapping multilateral donors. Late 2008, the CRS listed 263 international organisations to which core contributions count as official development assistance (ODA); this count includes public-private funds such as the Global Environment Facility (GEF) and Global Fund to Fight AIDS, TB, and Malaria (GFATM). The OECD (2008) study covered 229 agencies of the 242 that were on the list in 2007, since no information-not even a website-could be found for the other 13 organisations.

_________________________
The multilateral donors fall into three broad categories, according to CRS classification: • 47 UN agencies, funds and commissions, • the IMF (2 trust funds) plus the multilateral agenccies (5 bodies IDA, IFC, IBRD, MIGA, more than 1000 trust funds) of the World Bank, 12 regional development banks and funds, • other, which comprises NGOs, global funds and the European Commission with four bodies.
Add to this the 23 DAC donor governments with each a varying number of extending agencies, 35 international non-governmental organisations, five main public-private partnerships, and you get a rough idea of the 'old' donor cartel. The cartel has recently been complemented by a growing number of non-traditional donors, from China and other emerging markets, often in the context of a scramble for energy extraction rights (Paulo and Reisen, 2010). Figure 1 shows the number of organisations founded each decade by their main sector of activity. Early multilateral organisations are found in cross-border trade and communications, with the creation of the International Communications Union (1865) and the Universal Postal Union (1874). The bulk of the organisations/funds have been created since 1945. Only 15 of the organisations existed in 1940. In the 1940s, 15 more organisations were founded, following the creation of the Bretton Woods institutions in 1944 and the UN in 1945. The 1960s and 1970s saw an explosion of new agencies, with the creation of major bodies such as the UNDP and African and Asian Development Banks in the 1960s and 10 environmental (including UNEP and Habitat) and 10 agricultural research bodies in the 1970s. The 1980s and 1990s saw a rapid growth in agencies addressing governance and societal issues, most of them relatively small organisations. The health sector is often cited as highly fragmented, which may be in part due to the growing number of non-official actors. As far as official bodies are concerned, www.economics-ejournal.org there are now 34 health and humanitarian agencies on the list, almost half of them created since 1990. Table 1 provides information for country programmable ODA, the multilateral donor's share in total CPA, and respective staff numbers 4 , which excludes  Babb and Buira (2005, p. 59) on the Bretton Woods sisters:"Founded at the end of World War II to help lay the foundations of a new era of stability and prosperity, the World Bank and the International Monetary Fund (IMF) are widely viewed as having evolved in ways that would have surprised their founders. A term that has gained popularity among World Bank and IMF critics is "mission creep," or the systematic shifting of organizational activities away from original mandates."

_________________________
The so-called Malan Report 6 , a Fund/Bank-sponsored "Report of the External Review Committee on Bank-Fund Collaboration", released on 27 th February 2007, asked the International Monetary Fund to stop offering long-term finance to developing countries, leaving the World Bank to be the global development agency, as its work overlapped with the World Bank and that it should refocus its efforts on macroeconomic areas where it had greater expertise. Further, it admonished that there is currently no robust dialogue between the Bank and the Fund as they consider their future strategies and the implications this may have for how they work together. In particular, bank staff expressed concerns over directions in the Fund's medium-term strategy, particularly the role of the Fund in _________________________ 5 DAC member countries also channel large amounts through the multilateral system that are earmarked either by sector, theme, country or regionreferred to as multi-bi or non-core aid (USD 11 bn in 2006). These funds are reported as bilateral ODA in DAC Statistics. 6 To be sure, the Malan Report also pointed to many examples of good collaboration between the Bank and the Fund, such the Financial Sector Assessment Program (FSAP), the Heavily Indebted Poor Countries (HIPC) Initiative, and the debt sustainability analysis and framework.
www.economics-ejournal.org low-income countries and the division of responsibilities over financial sector issues.
Weary of agency patronage that links the two organisations to different ministries in most donor countries, the External Review Committee insisted that this should involve a meeting of 24 Governors, not 48. Each country and constituency should determine whether they will be represented at this joint meeting by their Bank or Fund Governor (where they differ). "This alone may encourage greater collaboration in capitals, particularly between Finance Ministries/Central Banks and Aid/Development Ministries" (IMF 2007, p. 8). Table 2 tries to juxtapose original mandates and information derived from recent mission statements. Substantial overlap becomes visible as the International Monetary Fund has encroached upon the fields originally occupied by the multilateral development banks, while the World Bank has tried to shift activities into areas originally dealt with by the UN system. So while the UN and the Bretton Woods Institutions were established with the intention that they would be complementary, their mandate shift has led to competition, overlap and duplication 7 .
The UN Development System has attracted considerable attention for its overlapping roles and mandates (OECD/DAC 2005). Three agencies in Rome are concerned with food security: FAO, WFP and IFAD; two UN organisations deal with health services for youngsters and young women: UNFPA and UNICEF; UNDP has three service lines related to AIDS, notwithstanding the UN organisation created to deal with Aids, UNAIDS; and the environment sector is taken care of by three UN bodies: UNDP, UNEP and UNIDO. Table 3 presents a notion of duplication and overlap in the area of 'aid for trade' 8 . _________________________ Source: Own tabulation based on mission statements found in annual reports.
The Report of the UN Secretary-General's High-Level Panel on UN systemwide coherence in the areas of development, humanitarian assistance and the environment (UN-Delivering as One), released on November 9, 2006, recommended "that the UN Secretary-General, the President of the World Bank and Executive Director of the International Monetary Fund set up a process to review, update and conclude formal agreements on their respective roles and relations at the global and country level". Two years later, Vatterodt (2008) finds that progress has been very limited: UN organisations still had not conducted an analysis of their comparative advantages and failed to define their division of labour on the basis of the results. To be sure, this finding would not surprise Dame www.economics-ejournal.org Anstee 9 . If the UN is unable to overcome its fragmentation, so Chataîgnier (2008), it will not overcome its marginalisation in the persuit of global public goods, to the satisfaction of unilateralists. Eight Millennium Development Goals (MDGs) were unanimously agreed by UN member nations in 2000. The Goals seek to focus international development efforts on the reduction of extreme poverty, the provision of universal primary education, combating key threats to health, environmental sustainability and a fairer international trading and financial system. United Kingdom's National Audit The reason lies in the entrenched vested interests of governments and of UN organizations and agencies, which saw their national, bureaucratic and personal fiefdoms threatened by the proposed changes. Those same forces will militate against the implementation of these latest proposals, and of any major reforms, unless there is a concerted effort by key governments of both developed and developing countries to generate the collective political will and commitment to see them through, including a radical change in the selection of the top management of the UN system".
www.economics-ejournal.org Office (2005) has conducted some simple exercises in an attempt to map the relationship between multilaterals and the MDGs. For each MDG and other thematic areas not covered by the MDGs (human rights, and humanitarian emergencies and conflict), Table 4 shows the key multilateral institutions which have stated objectives for achieving them. This analysis is based on the www.economics-ejournal.org multilaterals' own corporate information. The multilateral duplication and overlap in serving the MDGs is striking, costly and inefficient. Note also that the table does not reveal duplication and overlap within multilaterals, which can be striking 10 .

Explanations for the Rising Multilateral Donor Profilation
Explanations for the rising multilateral donor proliferation can be grouped as normative and positive. Normative explanations mostly center around the need to supply global public goods (Kaul et al. 2003) such as • climate change; food, water and energy shortages; global health problems • terrorism, ethnic conflict, and social fragmentation; and • the global economic order, financial instability, threat of protectionism, and job insecurity.
The switch from a unipolar world and the proliferation of now sovereign states can be taken as further normative explanations to explain multilateral donor proliferation. The global spillover of diseases, conflict, and financial crises and the inability of single actors to appropriate fully the benefits of remedial actions justify the creation multilateral actors who can help avoid the undersupply of global public goods: • Unilateral action fails because each country has an incentive to under-reveal demand for a non-excludable good or to reduce expenditures when allies increase theirs (Olson and Zeckhauser 1966); • Fiscal illusion and majority voting explain inadequate finance for international cooperation (Jones 2006). A public choice analysis of voters' behavior yields _________________________ the prediction that finance for international cooperation will be inadequate to redress failure in the global economy. Voters' perceptions are distorted (fiscal illusion), and reliance on specific voting rules exacerbates the impact of such distortion (through majority voting bias). Voters systematically underestimate the benefits of expenditure on international programs, particularly by comparison with expenditure on domestic programs.
Positive political economy, in contrast, explains the growth of international organizations without justifying it. The growth of international organizations can be attributed to the self interested utility-maximizing behavior of rational politicians and civil servants, including international bureaucrats who have a vested interest in the expansion of their organization (Vaubel et al. 2007). A seasoned observers of the multilateral aid scene has expressed a similar perception. To quote Andrew Rogerson (2004, p. 3): "In 50 years of aid no major institution has exited the market through closure or merger, with considerably more in existence today than when the share of aid in GDP was a third larger. They overlap in many ways under a rhetoric of 'harmonisation', (prevalent since Monterrey in 2002) and which is, doubtless coincidentally, selfpreserving. The creation of some institutions was deliberate, as a result of perceived deficiencies in existing ones. The latter were nonetheless allowed to continue, and even grow in parallel. Yet others are often seen as having little impact but struggle on in a diminished form through patronage ties, inertia, non-transparent funding formulas, and by eschewing any controversy that could tip political opinion towards outright closure." To be sure, staff numbers have declined over recent years in the FAO, ICAO, ILO, UNESCO and WHO (Vaubel et al. 2007). Still, Rogerson's view is supported by principal-agent theory as multilateral agencies pose a variety of two-stage principal-agent problems. Just as governments are supposed to be agents of their (median) voters, international organizations are appointed as agents of their member governments. Since both agents are only imperfectly controlled by their principals, there is a two-stage principal-agent problem. The principal-agent problem may have intensified over recent decades with repect to multilateral donors. Nielson and Tierney (2003) propose an agency theory of international organizations, addressing two persistent problems at international organizations: www.economics-ejournal.org i) long delegation chains from member governments (the principals) to the officers (so-called errant or runaway agents) who follow their own interests in maximising budget control and power; and ii) the common agency problem as member governments of an IO typically comprise a collective principal, and they must overcome a host of familiar collective-action problems before and while they delegate authority to an agent.
Moreover, the number of countries has risen in the wake of the breakdown of the former Soviet bloc, hence the common agency problem has intensified. As noted by Olson (1965) in his seminal study, agency slippage has a tendency to increase with the number of principals. Thus, bureaucratic inefficiency in international organizations is likely to rise with the number of member states. As the number of member states grows, the financing share of each member state and hence its share in the revenue from controlling the international bureaucracy decline. The governments, the media and the citizens of the member states lose interest in monitoring the performance of the international agent.
As the asymmetric information problem has increased with the number of countries and top-level development commitments, bureaucrats have taken advantage in requesting greater annual budgets, as suggested by the economic theory of bureaucracy; such budget appropriation is facilitated when there is no accountability in terms of clearly quantifiable and attributable results.
The average percentage ratio of salary to development financing is 2 for bilateral aid agencies, while it is 8 for multilateral donor agencies (Easterly and Putze 2008). The rise of staff in multilateral agencies makes it increasingly difficult, if not impossible, for bilateral agencies to monitor and control the biggest multilaterals, such as the World Bank. Vaubel, et al. (2007) analyse 27 international organizations in the years 1950-2001. During that period, staff increased at a compound average rate of 3.2% per annum, while the number of member states rose by only 2.5%. This may indicate a rise in bureaucratic inefficiency which takes the form of excess inputs, especially staff, as rising monitoring cost for the principals-the donor governments-have encouraged agency slippage by multilateral donors.
Much of the economic literature aims at rationalizing multilateral institutions relative to bilateral donors. Bulow et al. (1992) had explained a supposed superior enforcement capacity by multilateral development banks with de jure seniority (preferred creditor status), at the expense of bilateral creditors. The hypothesis was www.economics-ejournal.org based on evidence that IFIs extracted repayments better than bilaterals during the 1980s debt crisis. Note that the emrgence of Petro lenders and of China as new donors may have undermined discipline and reduced recipients' demand for multilateral lending. Cohen, Jacquet and Reisen (2006) find more defensive lending during the 1990s with multilateral than with bilateral lenders. This finding undermines the tale of the better extraction technology of multilateral donors. Rodrik (1995) postulates two multilateral advantages. First, since information on the quality of investment environments in different countries is in many ways a collective good, multilateral agencies are in a better position to internalize the externalities that may arise. Second, as long as multilateral agencies retain some degree of autonomy from the governments that own them, their interaction with recipient countries, while official in nature, can remain less politicized than intergovernmental links. This in turn endows multilateral agencies with an advantage in the exercise of conditionality. However, the collective-action/monitoring argument advanced by Rodrik would militate for concentration of multilateral financing mechanisms and organizations, not for the multilateral sprawl we are witnessing.
So it it does not surprise that Rajan and Subramanian (2005) find that multilateral aid is not more or less effective than bilateral, despite the presumption that multilateral aid is less tied, less explicitly "political", and better supported by conditionality than bilateral aid and should therefore have a different impact.

Towards Accountability and Efficient Assignment
Multilateral development finance has become too complex for both donor budgets and recipients' administrations. The proliferation of donors on the ground entails high transaction costs for all recipient countries 11 . Recipient-country administrations suffer from this complex system, overburdened by the number of interlocutors. Knack and Rahman (2007) analyse the impact of donor fragmentation on the quality of government bureaucracy in aid-recipient countries _________________________ 11 To be sure, some overlap between multilaterals may be useful. A degree of competition may create diversity in policy advice and service delivery and may add to the stability of aid flows. That said, however, a more harmonised approach will have a major role to play in raising the standards of aid delivery.
www.economics-ejournal.org and find that donor fragmentation leads to an erosion of bureaucratic quality. Competition among multilateral aid donors is therefore no solution for the supply and delivery of aid, a global public good with important externalities that tend to create natural monopolies.
Several steps are required to make progress toward a more accountable and efficient system of multilateral aid: • A prerequisite for effective ownership and efficient aid delivery, at the core of the Paris Declaration on Aid Effectiveness, is to map the rising complexity of multilateral development finance, to help identify areas for consolidation, address fragmentation and poor co-ordination at country level, and help identify comparative advantages for institutional role assignments among multilateral agencies. Such mapping identifies overlaps-leading to reduction of multilateral remit or proposals for consolidation; rivalries-leading to clarification of roles; and absences of co-ordination-leading to the design and implementation of co-ordinating structure. The OECD secretariat has made an important step toward mapping the multilateral donors (OECD/DAC 2008), but it will need to better informed about full staff numbers, administrative cost and aid allocation-by the multilateral agencies themselves 12 . Only few multilateral agencies presently report activity level data to the DAC secretariat, though not the all important UN and World Bank. • Many multilaterals claim to be working on the MDGs, but will they be held accountable when the MDGs go unmet? 13 Developing quantitative and qualitiative measures of multilaterals' contributions would be a useful step in promoting accountability. This would provide a basis to specialise multilateral _________________________ 12 The OECD (2008) finds the greatest opportunities for the multilateral agencies to concentrate their aid is in the 35 countries, where 9 or more multilateral agencies are in the long tail providing cumulatively less than 10% of a country's total aid. 13 Interactions between MDGs and responsabilities are highly complex so that traceability will be a hard problem to come by. Add to this the accountability that lies with the recipient countries themselves. However, both these arguments render any success claims by development agencies and multilateral donors pointless. And as long as these organisations continue to claim success, we must instill some degree of accountability. The MDGs at least have the advantage to break down into targets so that may actually facilitate traceability; moreover, they are timebound which opens the opportunity to try different organisational responsabilities over long time spans.
www.economics-ejournal.org agencies in line with the Tinbergen Rule 14 along the MDGs (i.e. goals and/or targets). As shown recently by Bourgignon et al. (2008), the correlation across MGD achievements is close to zero. On what basis would the MDG assignment to the various multilateral agencies be made? Formally, in a twotarget two-instrument (agency) context, we ask how the two instruments (agencies) must be varied conceptually to achieve the same change in one of the MDG targets. Then, given the required changes in the instruments, we need to determine how the other target responds to those same changes in the instruments. The agency that produces the largest absolute change in the other target is the one that ought to be assigned to that MDG target.
To be sure, these recommendations are far from producing operational instruments that would foster division of labour and accountability among the multilateral donor organisations. Staff selection, monitoring, procedural checks, and incentive-compatible contracts with managers of multilateral donors remain important workhorses toward less redundancy and more concentration in aid delivery.
Realistically, reform must start from outside, as vested interests in agency survival are strong. To make advances in streamlining the current (non-) system, existing circles of institutional patronage need to be broken. This requires a summit-level initiative that goes beyond the level of ministers (who are likely to defend 'their' international organisation). The many global tasks confronting world political leaders today-such as poverty reduction, global health, education for all, a clean environment-cannot be solved when yesterday's institutions duplicate with new players. New approaches to global governance do not only require a more inclusive and representative, but also a simplified and accountable system of multilateral development finance. 14 Tinbergen (1952) advanced the important principle that if governments aimed at n independent targets of policy they should also have n effective and unbounded instruments of policy if the targets are all to be met. If governments had three targets but four instruments, one instrument would be 'redundant' since only three need to be manipulated to achieve the targets.