International Aid for the Environment: Lessons from Economic Development Assistance

                                                                                                                                                                           

 

International Aid for the Environment:

Lessons from Economic Development Assistance

 

David Fairman and Michael Ross

94-03                                                   April 1994

 

 

CITATION AND REPRODUCTION

 

This document appears as Discussion Paper 94-03 of the Center for Science and International Affairs and as contribution E-94-03 to the Center's Environment and Natural Resources Program. CSIA Discussion papers are works in progress. Comments are welcome and may be directed to the author in care of the Center.

 

This paper may be cited as: David Fairman and Michael Ross. "International Aid for the Environment: Lessons from Economic Development Assistance." CSIA Discussion Paper 94-03, Kennedy School of Government, Harvard University, April 1994.

 

The views expressed in this paper are those of the authors and publication does not imply their endorsement by CSIA and Harvard University. This paper may be reproduced for personal and classroom use. Any other reproduction is not permitted without written permission of the Center for Science and International Affairs, Publications, 79 JFK Street, Cambridge, MA 02138, telephone (617) 495-1351 or telefax (617) 495-1635.

 

International Aid for the Environment:
Lessons from Economic Development Assistance

David Fairman, Harvard Center for International Affairs
Michael Ross, Princeton University

 

Executive Summary

This paper reviews lessons from the fifty-year history of international aid for economic development, and analyzes their implications for aid programs whose primary goal is environmental protection. We draw lessons from five issue areas: policy dialogue, conditionality, capacity building, aid agency comparative advantage, and inter-agency coordination.

Our major findings are

 

• The level of recipient political commitment has a dramatic impact on aid effectiveness, yet attempts by outside donors to induce this commitment where it does not already exist have generally failed. Determining a recipient's "true" level of commitment in advance of a program is particularly difficult, since recipients can easily feign commitment, and donors have a limited ability to monitor compliance. The best solution to this problem may be to insist that prospective recipients demonstrate their commitment to environmental programs by implementing large (and ideally, hard-to-reverse) portions of any aid-related agreements before funds are disbursed.
• Recipients' capacity to analyze, implement, and administer aid programs is a key variable in program success, but efforts by outside donors to foster these capacities have often failed. More research is needed on the determinants of effective capacity-building programs. It is clear, however, that environmental assistance programs have not yet taken seriously the painful experiences of development programs with institution building efforts.

• Donors with strong analytic capabilities, detailed local knowledge and long-term institutional ties to recipient counterpart agencies usually have greater influence on recipients than donors who seek to use aid volume and/or conditionality as a substitute for these capabilities and assets. Agencies which specialize in a small number of countries and sectors, and which target their interventions at a subset of institutional actors (e.g. ministers, sectoral agencies, local governments) may have a disproportionate influence in their areas of specialization. On the other hand, not enough has been done to integrate multiple agencies, project, sectoral and macroeconomic policy goals in the policy dialogue.
• It has been difficult for economic development aid institutions to learn from experience. Short-term organizational and political pressures have prevented donor and recipient policy makers from reforming development assistance policies and programs, despite the fact that many of these reforms would be in the long-term interest of both donors and recipients.
• It will be difficult to transfer lessons from economic development assistance to the new field of international assistance for environmental protection. Compared to economic assistance, environmental assistance generally requires more sophisticated political coalition building, greater interdisciplinary and interagency integration, and a longer time-scale for decision analysis. Two factors may, however, help policy makers and institutions learn more effectively in the field of environmental aid than they have in economic aid. First, recent work in the sub-discipline of environmental economics offers the potential for integrating many environmental and economic criteria in a common framework. Second, the uncertainties associated with environmental management may encourage more innovative and experimental approaches, and a greater commitment to self-evaluation and institutional learning.

 

Introduction

 

The purpose of this paper is to see what "lessons" can be drawn from a half century of experience with international aid for economic development, and to suggest possible applications of these lessons to aid programs whose primary goal is protection of the environment.* The scope of our research has been constrained by the limitations of the existing literature on international aid— which principally concerns the $500 billion in development assistance transferred to Third World states over the last forty years by the bilateral donors of the OECD, and by such major multilateral institutions as the World Bank, the IMF, and various UN agencies. While some of the technical lessons in economic development assistance do not apply to environmental aid, many of the strategic lessons appear highly relevant.

Our research comes at a time of considerable change in the world of international assistance. The end of the Cold War has brought about two major shifts: a redirection of development aid from the Third World to eastern Europe and the states of the former Soviet Union; and a search among OECD donors for a new, non-strategic rationale for assisting developing countries. This search has led donors to consider ways to use aid to promote sustainable development, democratization, and respect for human rights. At the same time, a gradual learning process among donors has led to more subtle changes, including a greater proportion of aid channeled through multilateral institutions, and a diminished proportion of aid "tied" to the purchase of products in the donor state.

The major Western donors have been reluctant to boost their environmental aid or to embark on ambitious new programs. The UN Conference on Environment and Development (UNCED) produced virtually no new commitments of funds from donors. The response of donors to the "national environmental action plans" of many developing states has been under-whelming; the World Bank's bid for a "green increment" to its MA X replenishment was rejected; and the Global Environment Facility is taking a modest project-level focus. The constraints on funding make it all the more important to find ways to improve the quality of the limited environmental assistance programs that donors are prepared to support.

 

The following sections review lessons from development assistance in five areas. We have not attempted to present technical findings on the relative merits of alternative economic development policies, or findings on the details of project and program management. Instead, we analyze major strategic questions which both donors and recipients must confront in the design of aid programs. Each section presents an overview of the issue, an analysis of the extent to which donors and recipients have changed their approaches to the issue over time, and a comment on the implications of the issue for environmental aid. We conclude with a summary of our findings, and an assessment of the potential for donors and recipients to apply lessons from development assistance to the design and implementation of environmental aid programs.

 

Policy Dialogue

Policy dialogue is the process by which an aid recipient and one or more donors negotiate changes in recipient government policies. Policy dialogue is not a new phenomenon in development assistance, although it has become much more prominent since the beginning of the debt crisis in the early 1980s and the expansion of structural adjustment lending by the IMF and the World Bank (Pearson Commission, 1969, pp. 127-133; Cassen, 1986, p. 69).1

During the 1980s, donors formalized the process of dialogue in many highly indebted countries. By making explicit the policy changes they expect recipients to make in exchange for receiving a specified amount of additional aid, donors have created a form of policy dialogue commonly known as "conditionality." The following section discusses the effectiveness of conditionality as a tool for promoting policy change. This section reviews the process of policy dialogue more generally.

Several factors determine the effectiveness of policy dialogue as a tool for promoting economically appropriate and politically sustainable policy reform. On the recipient side, political commitment, analytic and negotiating capacity increase the chances that policy reforms will be well-conceived and politically sustainable. On the donor side, depth of country experience, analytic capacity and long-term resource commitments increase the probability that policy advice will be useful and that donor promises to support policy reforms will be credible.

 

Donor Leverage and Recipient Commitment

Policy dialogue assumes both some shared interest in policy change and conflict over its form, pace and magnitude. Dialogue may be more or less adversarial, running the spectrum from thinly veiled coercion to full-fledged cooperation. In most cases, dialogue involves a mixture of persuasion, bargaining and coercion. The degree of conflict depends on the initial disparity between donor and recipient viewpoints, the analytic and negotiation skills each can bring to bear to change the views of the other, and the leverage (political and economic incentives and sanctions) each can exert to secure the other's agreement where differences persist (Jay and Michalopoulos, 1989b, p. 91).

Analysts and practitioners of policy dialogue generally agree that substantial donor leverage is neither a necessary nor a sufficient condition to induce recipient policy reform (see e.g. Mosley et al,, 1991, Chs. 4 and 5). The use of leverage may even backfire if it alienates the recipient government or causes the government to lose credibility and legitimacy in the eyes of powerful domestic constituencies (Berg, 1991, pp. 219-220). Donor insistence on ill-considered policy reforms has often led to disappointment and backlash.2

On the other hand, most analysts agree that recipient political commitment to policy reform is essential for success. Several characteristics of economic policy reforms make them high political risks. Reforms tend to impose immediate and certain costs on well-organized, well-informed interests (e.g. public enterprises, public sector workers and importers). The benefits of reform tend to come later, with less certainty, to actors who are less well-organized, less informed, and with less direct political access or influence (e.g. consumers. taxpayers, exporters) (Bates. 1994, p.31).

Domestic leaders contemplating the political risks of reform are rarely swayed by persuasion or coercion from external actors, except in cases where donors have exceptional economic or political leverage (Mosley et al., 1991, pp. 162-163). Instead, commitment to reform seems to depend on a conjuncture of one or more of the following factors: economic crisis, government or regime change, belief by decision-makers that current policies are both economically and politically unsustainable over the long term and that reforms will yield benefits fast enough to generate political support, a discredited and/or disorganized opposition, a tight-knit and influential team of reformist economists, and enough control of central and line ministries to push structural changes through state bureaucracies (Nelson, 1989, p. 12; Williamson and Haggard, 1994).

As Williamson and Haggard (1994) point out, none of these factors appears to be either necessary or sufficient to induce reform. Their relative weights are also highly context-specific. For example, the definition of crisis may vary widely across countries, and the impact of non-governmental supporters and opponents of reform may be greater in states with more opportunities for political participation. On the other hand, some weighted subset of these factors tends to be present in most reform cases. For example, Mexico did not experience a change of regime or government in 1982, nor was there widespread support for reform. Nevertheless, its economic crisis was extremely severe, and the leaders of the ruling party used their highly centralized control of authority and resources to overcome opposition despite the initial lack of public or elite consensus on the need for reform. Further, while the reformers succeeded in implementing a set of macroeconomic stabilization measures, they had to spend several years negotiating intensely with interest groups and winning the battle for public opinion before they could implement structural reforms such as privatization and financial liberalization (Cordoba, 1994, 277-283).

Aid agencies may only be able to influence political commitment at the margin, but they can perform several useful political functions. First, they can provide technical, financial and political support to help pro-reform economists in finance and trade ministries make their case to political decision-makers. Second, they can finance programs or projects to compensate groups which stand to lose from reform. This compensation need not replace one economic distortion with another. For example, it may be possible to compensate farmers for a reduction in fertilizer subsidies by undertaking an economically justified rural road-building program (Mosley et al., 1991, p.307). Third, they can improve their own credibility and leverage by reducing the volume of aid to recipients whose policies are glaringly inefficient and inequitable, and to recipients who promise reforms but consistently fail to implement them.3

In practice, aid agencies seem to have performed the political roles with mixed success. Case studies of reform in a number of developing countries suggest that aid agency economists have used their privileged access to senior political decision-makers to support reformers, but there is little evidence that aid agency advocacy has actually changed the minds of political leaders who were not already committed to reform. At best, agency staff may have encouraged political leaders to act sooner or undertake more comprehensive reforms than they might otherwise have undertaken (Mosley et al., 1991, Ch. 5; Williamson, ed., 1994; Kahler, 1992). Agencies seem have done little to provide compensation to groups adversely affected by reform, although structural adjustment loans can understood in the broadest sense as a form of compensation to import-dependent groups during the adjustment process (Mosley et al., 1991, p.129). Finally, agencies have rarely been able to make credible threats to reduce the volume of lending to recipients which fail to undertake reforms (see the following sections on conditionality and coordination).

 

Recipient Analytic and Negotiating Capacity

To the extent that recipient political commitment reflects fundamental political goals and constraints, policy dialogue can only affect it at the margin. In contrast, where weaknesses in recipient analytic and negotiating capacity make policy dialogue difficult, it may be possible for donors to assist recipients via technical assistance and capacity-building activities.

During the 1980s, a number of attempts were made to strengthen recipient analytic capacity during structural and sectoral adjustment programs, with mixed results. Paul (1990, p. 53-54) notes that the World Bank gave relatively less attention to building policy analytic capacity than to operations-related technical skills. Bank staff found it difficult to determine the appropriate organizational location, staffing needs and audience for policy analysis. It was also difficult to measure the contribution of investments in analytic capacity to the specific and relatively short-term goals of a particular sectoral adjustment package. Paul argues for earlier, more intensive, and longer-term support for the development of sector-level policy analysis capability, both within government ministries and in non-government settings.

 

Donor Analysis and Resource Commitment

Studies of policy dialogue emphasize the need for donors to base their policy advice on a sophisticated understanding of local economic and political conditions, and the need for donors to make credible, long-term aid commitments in support of policy reforms.

The need for sophisticated donor analysis seems obvious, yet in practice donors frequently advocate very similar remedies (e.g. a package of devaluation, price decontrol and deflation) in very different domestic economic and political contexts. Self-serving donor interests in opening markets, ensuring debt service and employing expatriate consultants may explain some of the failures of donor analysis. The ideological underpinnings and standard operating procedures of key multilateral donors may be another source of bias in policy advice. Bilateral donors often follow the lead of the World Bank and IMF in policy dialogue. These institutions have been criticized for excessive ideological commitment to neoclassical economic solutions, and for emphasizing short-term lending and adjustment targets over longer-term goals and strategies (Cassen, 1986, pp. 74-85).

While analyses of policy dialogue often focus on the need for credible recipient commitment to policy reform, credible donor commitments to support policy reforms with project and program aid appear to be equally important. Cassen (1986, p. 96) argues that "the credibility of policy dialogue in general depends on donors' ability to make and sustain long-term aid commitments." A long-term approach can have political, economic and institutional benefits. Politically, it may take several years to strengthen and stabilize economic and political constituencies with an interest in reform, and to provide adequate compensation to groups adversely affected by reform. Economically, long-term assistance can provide a buffer against external shocks and slower-than-anticipated supply and demand responses to policy reforms. Institutionally, it may take several years of financial and technical support to build the institutional capacity necessary to implement reforms, particularly when the reforms require increased technical sophistication.

Long-term commitments by donors may encourage and stabilize reforms. On the other hand, they may also allow recipients who are not committed to reform to delay action, making reforms even more costly and painful at a later date (Stern, 1991, p. 3). Donors have attempted to guard against this problem by disbursing funds in tranches, making disbursement of later tranches conditional on successful implementation of initial reforms. The following section examines the effectiveness of conditional tranche disbursement strategies in more detail.

 

The Ectent of Learning in Policy Dialogue

The donor community appears to have learned several useful lessons about policy dialogue from its long-term experience with program aid and from the intensive experience of structural adjustment lending in the 1980s. Short-term interests in improving the effectiveness of structural adjustment programs may have increased attention to the process of policy dialogue, but the analyses and reviews produced by academics and practitioners have gone beyond the specifics of structural adjustment to address broader questions about donor-recipient interaction at the level of policy goal-setting and strategy formulation.

Donors and international agencies have summarized the lessons learned on policy dialogue in a recent OECD Development Assistance Committee document, MAC Principles for Effective Aid" (OECD, 1992a). Key principles relevant to policy dialogue include the following:
 

• recipients must take primary responsibility for setting policy priorities;
• developing countries need well-designed policies which coordinate domestic investment and international aid:
• aid agencies should help strengthen developing countries' capacity to design and implement policies and programs; they should station policy-oriented staff in-country;
• donors should cooperate with recipients and each other via Consultative Groups and Round Tables, using multilateral aid agencies (World Bank, UNDP and RAF) as facilitators;
• donor advice can only be credible if it is competent, well-informed about recipient conditions and constraints, and based on multilateral consensus.

 

While donors do appear to have learned important lessons about the conduct of policy dialogue, learning among recipient countries is harder to assess. There is less literature available from a developing country perspective, and the continuing donor emphasis on building policy analytic capacity among recipients suggests that donors still perceive significant problems in this area. The thirty-year case studies of aid to India, Korea, Turkey, Ghana and the Ivory Coast presented in Krueger et al. (1989) show significant variation in analytic and negotiating sophistication among recipients. It is not clear, however, how much learning has taken place over time in each country, or what effect recipient learning has had on the policy dialogue. Other factors, particularly recipient regime changes, donor interests, donor sensitivity to recipient concerns, and the evolution of development and aid theory (including various aid fads), appear to have had a more direct impact on the content of policy dialogue and its effectiveness.

 

Implications for Environmental Policy Dialogue

The major lessons of donor and recipient experience with economic policy dialogue may apply a fortiori to environmental policy dialogue. The need to strengthen both donor and recipient understanding of the interaction of complex systems, the need for recipient political commitment and donor economic commitment over the long term, and the importance of building constituencies for policy reform are all highly relevant considerations for those engaging in environmental policy dialogue.

The environmental policy community appears to be well aware of these issues, as evidenced by attempts to create integrated national environmental action plans as a basis for donor commitments, and by efforts to strengthen recipient capacity for environmental policy-making and implementation (World Bank, 1993; UNDP, 1992). Though the track record of these efforts to date is mixed at best, policy dialogue is playing a role in increasing awareness of, interest in and capacity to implement pro-environment reforms (Arensberg, 1992).

Recipient political commitment is probably the key constraint on the implementation of pro-environment policy reforms. Long lags between human activities and environmental impacts and uncertainty about causal linkages make environmental "crises" difficult to recognize or explain. Lags and uncertainties also exacerbate the "short-term, concentrated cost, long-term, diffuse benefit" problems which economic reformers face. The relatively low status and power of most environmental ministries compared to finance ministries is a further handicap to environmental advocacy in developing countries.

On the other hand, most developing countries remain highly dependent on their domestic natural resources for subsistence and economic development. The poor in rural areas often face acute trade-offs between the short-term demands of subsistence and the long-term need to protect their environmental resource base, but they also have a strong interest in supporting policy packages which mitigate these trade-offs. For example, many pastoralist groups maintain unsustainably large livestock herds because livestock are a store of capital. Providing pastoralists with access to basic banking services can lead to reductions in herd size and increased incomes (Dasgupta and Mahler, 1994, p.6). The more general point is that when developing country governments are interested in creating sustainable development strategies for the rural poor, they should be able to do so, often with net economic as well as environmental gains (WRI, 1992, Ch.3). Unfortunately, the fact that the rural poor are the group with the most to gain from efforts to protect the natural resource base leads directly to the conclusion that there are very substantial political impediments to the policy reforms that would benefit them. The obstacles to political participation by the rural poor are severe, and the changes in land tenure, allocation of credit and technical assistance, taxes and subsidies that would be required to promote sustainable rural development confront vested interests at least as strong as those opposed to more conventional forms of economic policy reform (Ascher and Healy, 1990, Ch. 2).

As in economic development assistance, the key question is whether external aid can substantially increase political support for environmental protection within developing country governments and among key social actors. If aid agencies can use some combination of persuasion, compensation and coercion strategies to affect economic development priorities at the margin, they should also be able to affect environmental priorities. The major problem with the analogy, as noted in our introduction, is that donor governments and aid agencies generally have much less interest in assisting developing countries with environmental management than with economic development. Although donors have increasingly embraced the rhetoric of sustainable development, it is not clear how much their words represent a real change in beliefs and values, and how much they are a symbolic response to political pressure from advocacy coalitions of developed country environmental NGOs. In some cases, however, networks of developed country environmental NGOs and developing country NGOs representing the interests of the poor have been able to put pressure on donors. aid agencies and recipients to reorient environment and development polices, and to hold them to their word by monitoring policy implementation. Where NGO networks can extend build alliances with government agencies, elected politicians and businesses which stand to gain from a change in development policies (e.g. renewable energy producers), they may become a powerful political force for promoting sustainable development (Rich, 1990; WRI, 1992, pp226-232).

Finally, even with greater awareness, political and financial commitment on the part of both donors and recipients, environmental policy dialogue conducted in isolation from economic policy dialogue is not likely to improve or even maintain environmental quality in most developing countries. Greater integration of environmental and economic policy issues in policy dialogue, and a more serious search for complementarity in economic and environmental policies, are prerequisites for success in countries which depend heavily on natural resources for subsistence and growth. Without integration and a search for "win-win" strategies, narrowly-focused environmental policy dialogue is likely to be undercut by economic and social policies working at cross purposes with environmental goals (World Bank, 1992, Ch.3).
 

 Conditionality

Conditionality can be defined as the exchange of financial assistance for policy reforms. There is a large literature on the most common use of conditionality, the structural adjustment and sectoral adjustment loans offered by the World Bank to developing countries in exchange for a variety of financial reforms. There is a much smaller literature on— and much less real world experience with— the use of conditionality for the pursuit of non-financial goals.

Conditionality can take two forms. The more common type, which has been used in the case of structural adjustment, is "policy-based lending," in which donors disburse funds in exchange for the recipient's adoption of a set of prearranged policy changes. The second type, which has recently been adopted in one form or another by most OECD countries, is "allocative conditionality," in which donors distribute varying amounts of aid to countries based on their adherence to specified standards, but (in theory) do not attempt to bargain with them to bring about reforms.

This section reviews the findings of recent scholarship on the effectiveness of each type of conditionality; examines the compliance and enforcement problems they face; and makes inferences about the applicability of these lessons to environmental assistance.

 

Policy-based Lending

The literature on the effectiveness of policy-based lending is extensive. In theory, policy-based loans may be linked to either financial or non-financial criteria. In practice, their main focus has almost always been financial reform, though non-financial conditions have sometimes been added. 

Measuring the effectiveness of policy-based lending raises three methodological problems: selection effects (those countries willing to accept conditional aid are a self-selected group that is probably better prepared to enact reforms than non-applicants); the complications of establishing counterfactuals (i.e., what would have occurred in the absence of conditional aid?); and the ability of recipient governments to employ "false compliance," by enacting the specific conditions agreed to, while nullifying their impact with countervailing measures. Nonetheless, policy-based lending has been closely studied, and there is a good deal of evidence on the limits to its utility.

A number of scholars have conceptualized structural adjustment conditionality as a form of bargaining, since international financial institutions (IFIs) appear to be offering recipients economic assistance (in the form of concessional loans) in exchange for their promise to enact reforms (Mosley, 1987; Kahler, 1992; Schoultz, 1989). Some evidence appears to validate this conception; other evidence contradicts it. Taken as a whole, seven broad lessons emerge on its effectiveness in promoting policy reform. The first three lessons concern the internal capacity of IFIs to negotiate, monitor, and enforce conditionality agreements; the rest concern the limitations of conditionality as a policy tool.

First, IFIs have turned out to be poor bargainers and poor contract enforcers for a number of institutional reasons. The World Bank (the principal IFI in many cases of structural adjustment lending) has virtually no flexibility in the size of the loans it can offer to a recipient: the size is set far in advance and is known to both parties. Hence the Bank is unable to "buy" more reforms from potential recipients by offering them more money (Mosley et al., 1991, p. 115). Even when it can credibly make its lending program conditional, the Bank may not be able to state or measure its reform requirements in quantitative terms; qualitative conditions are much harder to enforce because of their intrinsic ambiguity. Most importantly, IFIs are under great institutional pressure to disburse loans, regardless of the compliance of recipients with the prearranged conditions. IFIs have an institutional interest in making loans, which are necessary to remain profitable. and IFI managers have a bureaucratic interest in spending the full amounts of their budgets. Hence, the larger the country and the more important its status as a borrower, the less able IF Is are to refrain from offering them loans while still spending their budgets and remaining solvent.

In addition, the World Bank has periodically been coerced into making loans to states that are poor credit risks but strategic allies of the Bank's major benefactor, the U. S. For these reasons, the most important borrowers are generally given the "softest" conditions, and lenders virtually never cancel a loan once disbursement has begun, no matter how poorly the recipient is complying with the conditions (Mosley et al., 1991, pp. 48, 71-72, 165).

Moreover, if a borrower has fulfilled at least 50 per cent of the conditions in its previous loan (and if it is an important borrower or has a close relationship with the Bank), the Bank is almost certain to offer it additional loans once the initial loan has been disbursed. There is now so much evidence of the Bank's laxity that borrowers no longer feel obliged to fulfill more than half the conditions they initially agree to (Mosley et al., 1991, pp. 170-171). The Bank's eagerness to disburse its funds has one additional consequence: it gives Bank officials an incentive to overestimate what a recipient can accomplish with a conditional loan. Since their counterparts in national ministries of finance share this incentive, they tend to collaborate in setting unrealistic targets for recipient countries, which they cannot meet under the best of circumstances (Mosley et al., 1991, p. 109). 

Second, because IFIs have limited capacity to monitor their agreements, there is a tradeoff between the number and complexity of the conditions in the agreement and the ability of lenders to enforce them. The World Bank's abundant bargaining power has turned many lending agreements into "Christmas trees." with every Bank officer determined to attach his or her own conditions. Agreements with many complex conditions are difficult to enforce for three reasons: they overburden the recipient government (particularly because the Bank has been unable to find ways to prioritize its conditions); they overburden the Bank, which has a finite ability to monitor compliance, and because many of the conditions have interactive effects, they make it difficult to distinguish between governments that are committed to reform and those that are covertly defecting from the agreement (Mosley et al., 1991, pp. 44-45).

Third, while the Bank has certainly "learned" from its past mistakes, it has often been unable to implement the new "lessons." Over the course of the 1980s, the Bank's own analyses stressed the need to impose fewer conditions on borrowers, to prioritize the conditions, to insist on full compliance before releasing the second (or third) installments of the loan, and to arrange for recipients to make more significant reforms before the loans are disbursed. Of these four lessons, the Bank has only been able to enact the last (Mosley et al., 1991, pp. 128-129). The institutional incentives to over-prescribe and under-enforce have thus far been intractable.

The fourth lesson is that compliance with a conditionality agreement has little to do with the recipient's type of government and a great deal to do with the characteristics of the issues at hand. The compliance of all types of governments is likely to be greater on issues with the following attributes:
 

a.         they are not politically sensitive (mandates to limit wages, for example, have been relatively unsuccessful) (Kahler, 1992, p. 97);

b.         they can be put in place by a small number of officials, and do not require the cooperation of large numbers of bureaucrats or agencies;

c.         they can be easily monitored;

d.         they can be accomplished quickly and in a single step (Nelson,1989);

e.         they do not require extensive institutional change to implement;

f.          they "break bottlenecks" by opening the way for further actions by the government or private actors; there is a strong technical consensus on their utility (Nelson and Eglinton, 1993, p. 69)

 

Fifth— and perhaps most surprisingly— there is little evidence that policy-based loans actually induce recipients to embark on economic reform programs. Borrower governments have usually been either firmly opposed or strongly committed to the type of reforms favored by the Bank, quite apart from the influence of conditionality. Even when there are internal battles between pro­- and anti-reform factions, the outcome is usually clear and has little to do with any enticements offered by IFIs. Governments that wish to obtain the loans and then "defect" by blocking reforms have been able to do so, partly by taking advantage of the Bank's limited enforcement capacity, and partly by engaging in "false compliance" by fulfilling the letter of their obligations but nullifying their impact with countervailing measures. The ease of defection, combined with the eagerness of the Bank to disburse loans, gives those governments lacking a strong commitment to reform few incentives to take painful measures.4

 

Sixth, despite the inability of IFIs to persuade recipients to undertake reforms they oppose, policy-based lending has played an important role in supporting governments already committed to reform by helping them to restructure their economic institutions and buffer themselves against political opponents. Bates and Krueger (1991) find that governments that implemented successful economic reforms had been able to increase the power and discretion of economic technocrats, shift the locus of economic policy-making from the representative to the executive branch of government, and find the resources to build a political constituency for the reforms. A conditional loan package can assist each of these tactics by making funds available and by forming close alliances with a ministry of finance or a central bank. Burgess (1993) contends that directives from IFIs can support governments committed to economic restructuring by helping them institutionalize reform programs, and by giving their public officials a more defensible rationale for taking politically unpopular measures. And Mosley, Harrigan, and Toye (1991, pp. 119-120, 160) find that IF Is can do three things to help recipients already committed to reform: use their resources to remove technical obstacles to reform; aid pro-reform coalitions by providing them with information and devising common strategies to circumvent the legislature or obstinate bureaucracies; and provide funds to buy out opponents to reform.

Still, while policy-based lending may help fortify like-minded technocrats, the prior commitment of the political elite to reform remains crucial: in both Zaire and the Philippines, autocrats temporarily empowered reform-minded technocrats while negotiating policy-based loans, then rescinded their authority once the loans had been approved (Kahler 1992, p.128).

The seventh and final lesson is that there are techniques which can and should be employed to force recipients to reveal their true preferences (i.e. their commitment or lack of commitment to reform) in advance of a policy-based loan. Having belatedly realized the importance (and the opacity) of recipient commitment, 1F Is now frequently insist that prospective borrowers make large "down payments" by enacting painful reforms in advance of any disbursement, to demonstrate their commitment to reform and reduce the chance of defection or false compliance (Kahler 1992, p. 117; Mosley et al., 1991, pp. 128-129). The form of conditionality discussed below— allocative conditionality— attempts to take this last lesson one step further.

 

Allocative Conditionality

Allocative conditionality— a strategy by which donors refrain from bargaining with recipients altogether, and instead allocate their funds according to the adherence of potential recipients with prescribed standards— has recently drawn a good deal of attention in the foreign aid community. Since 1989, most bilateral donors have announced that the allocation of their development assistance will now be based, in part, on the recipients' poverty-reduction programs, human rights standards, reductions in military spending, and progress towards "good governance" and democratic reforms. The European Bank for Reconstruction and Development (EBRD) goes even further, offering its services only to states that uphold human rights and the rule of law and have competitive elections. Scholarship on the utility of allocative conditionality as a policy tool is still, not surprisingly, quite preliminary (Ball, 1992; Lewis, 1993; Nelson with Eglinton, 1992; Nelson and Eglinton, 1993).

Nelson and Eglinton (1993) examine the impact of recent decisions by donors to reduce or suspend their aid to countries that egregiously violated human rights or had been "backsliding" on democratic reforms. The findings are mixed, but suggest that an aid cutoff was more likely to catalyze policy changes if the recipient had a strong, organized opposition that supported reforms (e.g., Guatemala) rather than a weak or divided opposition (e.g., Kenya and Malawi), or if the country was governed by an autocrat who demonstrated concern for the public welfare (e.g., Peru's Fujimori) rather than simply his own survival (e.g., Haiti's Cedras)--though the latter finding appears somewhat tautological.

There is insufficient data on the use of allocative conditionality to reward good behavior to draw any conclusions about its efficacy; the foreign aid budgets of most OECD countries have been too tightly constrained to permit much experimentation. Nonetheless, by announcing explicit standards for allocative conditionality, bilateral donors may have some impact on the expectations of current and potential aid recipients, who conceivably may alter their behavior in anticipation of the judgment of donors— an impact that would, however, be difficult to measure.

Nelson and Eglinton find that allocative conditionality has thus far "been implemented only partially and inconsistently." Donors have penalized a handful of extreme violators (Zaire, Myanmar, and Haiti) and rewarded a handful of countries undertaking democratic reforms (Zambia, Poland, Russia), but have failed to systematically shift aid levels over the full array of recipients (Nelson and Eglinton, 1993, p. 50). They also suggest that allocative conditionality will necessarily remain a crude and partially-implemented foreign policy tool, since bilateral donors have neither the administrative capacity to carefully gauge the progress of recipients towards a complex set of goals, nor the desire to forsake their prior commitments and interests in allocating foreign aid. In general, however, allocative conditionality faces far fewer implementation and compliance problems than policy-based lending, since under allocative conditionality, the donor rewards recipients for their prior behavior rather than entering into a contract with them to improve their future behavior.
 

Lessons for Environmental Assistance

The structural adjustment loans offered by international financial institutions to debt-ridden developing countries may seem to be a poor analogue to most types of environment transfers. They involve large amounts of money— often hundreds of millions, or even billions of dollars— in a single package; they are designed to foster far-reaching reforms that may impose significant short­-and medium-term costs on many sectors in the recipient countries; and unlike many environmental transfers, they are usually politically costly to the recipient, as they entail a large reduction in state institutions and reduced opportunities for patronage.

Nonetheless, the lessons drawn from structural adjustment conditionality may be germane for environmental assistance programs for at least four reasons. First, conditionality has been carefully studied and provides a wealth of data on the political dynamics of conditional financial transfers. Second, there may in the future be types of environmental transfers that more closely resemble structural adjustment. Third, the institutions that make financial transfers for the environment face many of the same constraints (and in fact, are often the same institutions) as those involved in structural adjustment lending. Finally, conditionality provides an opportunity to study the behavior of states and institutions under a set of extreme conditions, in which donors appear to have maximum leverage over recipients. By extrapolating from this extreme setting— in which the stakes are high for both lenders and recipients— it may be possible to gain insights into the utility of conditionality under more mundane circumstances.

While "green" reforms may be less distasteful to government leaders than structural adjustment, a decade of experience with conditionality suggests that even large environmental aid packages will be unable to coerce reluctant governments to implement policies they oppose— particularly when those policies are politically costly, cumbersome to implement, or difficult to monitor. On the other hand, when properly structured, conditional environmental transfers may help a "green" political elite overcome substantial technical and political obstacles to reform. Conditionality is paradoxically ineffective in provoking reform, but can be highly effective in nurturing it.

 

Capacity Building

Perhaps nothing in the field of development is as popular to promote or as difficult to accomplish as capacity building. For decades, development specialists have tried to foster what we now refer to as "capacity building" under the rubrics of institution building, institutional development, public sector management, and public administration. The great majority of what has been written on this topic is normative and concerned with practical management questions; relatively little evaluative work has been done. Still, there are at least six lessons to be gleaned from the handful of high quality studies.

The first is that capacity building has been the Achilles' heel of a great many development projects. Most World Bank projects can be broken down into a "hard investments" component (equipment and physical infrastructure) and an "institutional development" component (which includes the promotion of planning, accounting, maintenance, staff training, and other forms of technical assistance). Assessments by the Bank have found that the institutional components of programs fail about twice as often as the physical components; and that the most common obstacles to the successful implementation of projects are "managerial" or "institutional" problems (Israel 1987, p. 3). A review by USAID of its support for institutions found the results "about half positive and half negative"; the Canadian development agency found that two-thirds of the recipient country institutions it had assisted were unprepared for self-reliance at the end of the project (Cassen, 1986, p. 111). In fact, many development practitioners have found the problems of "institutional weakness" so intractable that they prefer to ignore the issue altogether and instead worry about more tangible problems.
 

Second, ex post evaluations of World Bank programs have found that the success of institutional development programs varies less by country than it does by sector, subsector, and activity. The most successful programs have been in industry, telecommunications, some utilities. and non-rural finance; the least successful in agriculture, education, and railways. Capacity building efforts on technical and financial issues have done relatively well; initiatives promoting interagency coordination, personnel management and training, maintenance of facilities and infrastructure, and sector-wide reforms have done relatively poorly (Uphoff, 1986, p. 196; Israel, 1987, pp. 19-20; Cassen, 1986, pp. 195-198). Good overall explanations for these patterns are lacking, though it seems clear that setting up effective and efficient institutions is much more difficult in rural than urban areas (Rondinelli 1986).

Third, in many aid projects there is a conflict between the short-term interest of donors in efficiently completing hard investments, and the long-term goal of fostering institutional development. Donors are usually anxious to identify problems in recipient countries and supply the necessary organizational skills and technologies (often originating in the donor's own country) to solve them. This may be the most efficient way to approach a discrete "development" problem in low-income countries. But it also may have adverse consequences for any capacity building efforts, since it tends to deprive local officials of the planning, organizational, and problem-solving skills they will need to maintain the project. The problem is exacerbated by the desire of donors to show their own taxpayers the tangible results of foreign aid programs as quickly as possible, and their interest in hiring their own citizens and firms to perform much of the work.

Fourth and related, the institutional components of development projects frequently fall apart once donors leave, often because the recipient country is unable to afford the recurrent costs of staffing and maintaining a project. Governments of low income countries typically find it easier to acquire funds for hard investments from international donors and financiers than to raise the operating costs to maintain the institutions surrounding the project once the donor leaves— commonly because the recipient must operate from a weak and politically volatile tax base and/or has poor macroeconomic policies. The scarcity of funds to maintain donor-initiated institutions means, on a practical level, that new bureaucracies cannot fund their operating costs and will deteriorate; and that roads, vehicles, and other infrastructure must operate well below capacity and over time may have to be abandoned (a ubiquitous phenomenon in Africa). Capacity Building efforts, if they are to be sustained, must include provisions for long-term recurrent financing (Duncan, 1985, pp. 136-137; Mosley et al., 1991, pp. 30-31).

Fifth the most important reason capacity building efforts fail in both low and middle income countries is the absence of a durable and bureaucratically widespread commitment by the recipient country. Donors and recipients commonly have differing views on the goals or structure of the institutions targeted by capacity building efforts; if the recipient fails to give the institution sufficient economic, political, and bureaucratic support once the donor leaves -- or fails to sustain that support over a period of decades -- the institution may well not survive (Israel, 1987, pp­40-41).

Finally, developing country NGOs are playing an increasingly important role in augmenting the capacities of their governments. The number and resources of southern NGOs vary widely: they are thought to be strongest in Asia and Latin America, weak in Africa, and virtually non-existent in the Middle East, North Africa, and northeastern India. Though the figures are unreliable, the number of NGOs in developing states appears to have tripled since 1977. In some states, NGOs have formed alliances with governments and have taken on quasi-governmental functions that governments themselves may be ill-equipped to perform, such as delivering services in rural areas and mobilizing resources at a local level. Southern NGOs appear to be increasingly successful in attracting financial support directly from the North (often from other NGOs), thus serving as conduits for capacity building efforts that bypass the state completely (Livernash, 1992; Dichter, 1988).

Implications for Environmental Transfers

Perhaps more than any previous form of development assistance, environmental aid transfers are aimed at capacity building; hence the lessons from previous efforts to foster "institution building," "institutional development," and "public sector management" are especially salient. Many environmental assistance projects treat capacity building far too casually, assuming that a short-term barrage of technical training programs will bring about. sturdy environmental institutions in low and middle-income countries. Environmental aid donors have paid far too little attention to the need for an enduring political commitment from the recipient government for successful capacity building; and scholars have thus far have been of little help. As a result, practitioners and scholars risk learning the lessons of earlier capacity building efforts in the most costly and frustrating way possible.

 

Comparative Advantages of Aid Agencies

A wide variety of aid agencies provide funds, technical assistance and capacity building services to aid recipients. Not all agencies are equally effective at performing these tasks, and some analysts have argued that differences in agency governance structures and lines of accountability constrain agency effectiveness. In particular, some have argued that there are systematic differences in the effectiveness of bilateral, multilateral and nongovernmental agencies. While there is some evidence to support these claims, distinctions at this level of generality may not be relevant for policy-making. Furthermore, recent studies suggest that differences among agencies within a category may be as important as differences between categories. The debate has evolved toward a more nuanced discussion of agencies' comparative advantages in specific contexts.

Bilateral and Multilateral Agencies

Bilateral agencies (e.g. USAID, SIDA, CIDA) are created by and responsible to individual donor governments. Multilateral agencies (e.g. the World Bank, IMF, regional development banks, UNDP and specialized UN agencies) are created by and responsible to a group of member states, usually including both donors and recipients.

Cassen (1986, pp. 268-269) and Jay and Michalopoulos (1989a) review the key issues in the bilateral-multilateral debate. Some analysts have criticized bilateral agencies for promoting projects and programs which serve their governments' short-term, commercial and political interests at the expense of recipients' long-term development needs. As evidence, they note that 1) donors channel approximately 70% of their aid through bilateral channels, 2) less bilateral than multilateral aid is allocated to the poorest countries, and 3) at least half of all bilateral aid is tied to donor exports. Others have praised bilateral agencies for their ability to promote development goals in the context of broader political, cultural and strategic relationships. Those who have criticized the short-term and self-serving bias of bilateral agencies have praised multilateral agencies for supporting recipients' long-term development goals, while others have criticized the multilaterals for their ideological biases, bureaucratic inefficiencies, and unwillingness to challenge recipient government policies.

Despite the intensity of this debate, it is not clear that donors can make short-term choices between bilateral and multilateral agencies to achieve particular aid goals in individual countries. Most donors do not have an institutionalized process for allocating aid at this level of specificity. Rather, they make incremental changes in the balance between their bilateral and multilateral aid budgets in annual aid appropriations. Aside from these appropriations decisions, most operational task assignments and resource commitments are negotiated among bilateral and multilateral agency representatives at the country and regional levels. In this setting, the agencies act as principals. Their allocation decisions are self-interested, raising the coordination problems discussed in the following section.

Even if donors had the capability to make detailed operational choices between bilateral and multilateral channels, several recent studies of agency effectiveness suggest that the bilateral-multilateral categorization may not be a useful framework for analysis. Variations in roles and comparative advantages are at least as great within these categories as between them.

A study of European bilateral aid policies by Arnold (1982) shows that Sweden and the Netherlands have a strong commitment to long-range development goals and basic human needs approaches; short-term commercial and political goals have little effect on aid decision-making. Britain and France, on the other hand, pursue political and commercial objectives along with humanitarian ones. A recent GAO report (GAO, 1993) on USAID acknowledges that anti-communism provided the driving rationale for US aid policy during the Cold War. It also points out, however, that the combination of multiple, vague and conflicting aid objectives and AID's decentralized management system allowed field missions to pursue a wide variety of goals, including basic human needs and long-term development.

Multilateral aid agency policies, structures and capacities vary equally widely. A well-researched Danish aid agency report (Danida, 1991) compares the effectiveness of eleven multilateral aid agencies in four countries. The report divides multilateral agencies into UN funds and programs (UNDP, UNSO, UNICEF), UN specialized agencies (FAO, ILO, WHO), regional development banks (AfDB, ADB), humanitarian agencies (UNHCR), the World Bank, and the EC aid program. It proposes ten potential advantages which an aid agency may have, and rates each subset of agencies in terms of its potential to achieve these advantages (p.7).5The report then analyzes the actual performance of each agency in each country, identifies particular strengths and weaknesses, and proposes "concentration" of each agency on the areas in which it appears to have an actual comparative advantage. For purposes of this discussion, the key point is that the category "multilateral agency" is not nearly as useful a guide to agency selection as a scheme which classifies individual multilateral agencies functionally and examines their individual, operational track records.

 

Nongovernmental agencies (in both donor and recipient countries) are created by and responsible to private, nonprofit groups, although they ma), also, work as contractors to bilateral and multilateral aid agencies. During the 1970s and 1980s, NGO memberships, capacities and activities expanded dramatically in both donor and recipient countries, Donor-based NGOs now contribute approximately $3 billion in annual ODA roughly five per