CONDITIONALITY

The concept of conditionality refers to ‘the promise or increase of aid in the case of compliance by a recipient with conditions set by a donor, or its withdrawal or reduction in the case of noncompliance’ (Frerks Reference Frerks2006: 21). Building on this definition, Frerks (ibid.: 16) notes that peace conditionality is aid that is used ‘as a lever to persuade conflicting parties to make peace, to implement a proposed peace accord, and to consolidate peace’. Thus, conditionality necessarily implies the presence of some form of pressure by donor agencies, whether the pressure is for political outcomes, such as elections, or economic outcomes, such as fiscal decentralisation.

Patterns of aid conditionality have changed quite dramatically, in scope and form, over the last several decades. The modern conception of aid developed in the mid-twentieth century, after 1945, with the Marshall Plan as ‘a powerful example and model for forging development aid relationships’ through the use of ‘explicit political and economic conditions’ to rebuild war-torn countries (ibid.: 5). By the 1980s, the World Bank, IMF, and many bilateral donors had begun regularly to impose aid conditionality aimed at inducing recipient country governments to put in place a series of macroeconomic policies that donors believed were essential for promoting economic growth.

The end of the Cold War resulted in the creation of a new, more comprehensive and coordinated aid regime among the world's major donor countries. Patterns of aid also incorporated an additional layer of conditionality that was aimed at promoting good governance and democracy (Suhrke & Buckmaster Reference Suhrke and Buckmaster2005: 16). This pattern carried over into internationally led post-conflict peacebuilding operations. Political liberalisation was often an explicit part of the peace agreement in these countries, and war had heavily damaged their economies. Thus, democratisation and economic recovery became inextricably bound up with peacebuilding.

WAR AND PEACE IN MOZAMBIQUE

The General Peace Accord (GPA), signed in Rome in October 1992, put an end to sixteen years of civil war in Mozambique. The war began a few years after the country gained independence from Portuguese colonial rule in 1975. At independence, Frelimo, the Front for the Liberation of Mozambique, which had fought an armed liberation struggle against the Portuguese, established a single-party regime in Mozambique. Frelimo initially enjoyed widespread support because of its leadership of the independence struggle, as well as its call for national unity, mirroring the post-colonial political experience of numerous other newly independent African states (Manning Reference Manning2002). The Frelimo government adopted a Marxist–Leninist platform in 1977, the same year that the rebel group Renamo launched its first attacks. Frelimo proceeded to enact ambitious reforms and nationalisation projects, which were largely disruptive to Mozambique's society and economy. These changes, along with regional and international factors, including the Cold War, helped to fuel Mozambique's civil conflict.

By 1988, both warring parties had begun to search for an alternative to continued conflict. Changes in the broader international context played a key role in bringing peace to Mozambique. The first moves towards negotiation to end the war coincided with a series of interrelated international events: the end of the Cold War; the adoption by major donors of democracy as a condition of aid; the toppling of authoritarian regimes in the Soviet sphere of influence; and the end of apartheid in neighbouring South Africa (Chan & Venancio Reference Chan and Venancio1998; Manning Reference Manning2002; Vines Reference Vines1991).

ANTECEDENTS OF EFFECTIVE CONDITIONALITY IN MOZAMBIQUE'S PEACE PROCESS

Conditionality, as we have noted, involves problems of information, cooperation, and coordination. This section describes how decades of donor assistance, first from governments sympathetic to Frelimo's socialist vision, and later from a broader spectrum of donors, created a group of major donors with substantive knowledge about both state capacity and the challenges the state would face in post-war reconstruction. These donors were able to identify and respond to shortcomings in the UN-led peace process as they arose. We argue that this experience facilitated a common understanding by these donors about what had to be done in order for the peace process to succeed.

Second, we argue that donor involvement during the war and drought that preceded the GPA facilitated donor cooperation and coordination. In the heat of the humanitarian emergency, donors had devised several coordination platforms that allowed them to overcome common collective action problems, as well as to circumvent more formal, but often dysfunctional, coordination mechanisms. These humanitarian emergencies showed Mozambique's donors the benefits of pragmatic, ad hoc coordination, although they also demonstrated the precariousness of such efforts (Ball & Barnes Reference Ball, Barnes, Forman and Patrick2000).

In addition, two other factors offered motivation for donors to overcome the usual barriers to coordination. First, the failure of the UN-led peace process in Angola in 1992 reinforced donor motivation, and international dedication, to ensure a successful transition to peace in Mozambique. Second, unlike many other countries where external actors were engaged in brokering peace, Mozambique was of little strategic political or economic importance for its major donors. Realistically, ‘Mozambique was never high on the priority list for any of the major world powers’ (Howard Reference Howard2008: 188). In the words of Andrea Bartoli (Reference Bartoli and Crocker1999: 188), one of the mediators in the country's peace process, ‘Mozambique was, in a way, marginal to the sociopolitical map of the early 1990s. No longer under the influence of a collapsing communist world, no longer a threat to a changing South Africa that was abandoning apartheid, Mozambique did not fall under the international spotlight.’ The knowledge and experience gained during the war, together with the strong desire for a success in Mozambique and the absence of divisive strategic resources, help explain the successful coordination among donors that was essential to the effective use of peace conditionalities.

MOTIVES AND OPPORTUNITIES FOR DONOR COORDINATION

As Boyce (Reference Boyce, Forman and Patrick2000: 378) has noted, ‘peace conditionality translates the grand bargain between the international community and the warring parties into a series of mini-bargains, which link disbursements to specific peace-building measures’. The effective use of peace conditionalities requires adequate information, as well as donor willingness and ability to cooperate on shared goals and coordinate their actions. Long donor experience in Mozambique went a considerable way towards providing information necessary for the formulation of goals and strategies, and provided precedents for donor cooperation on an ad hoc basis.

However, for donors – individually and collectively – to leverage their aid money in exchange for compliance on the myriad ‘mini-bargains’ that constituted the implementation of the General Peace Accord, coordination had to be considerably refined. The long-standing involvement of major donors in Mozambique translated into collective efforts to anticipate and address shortcomings in the peace plan as the implementation process unfolded. In this section, we examine why and how major donors were willing and able to do so. We then show in some detail how a group of major donors used an ad hoc coordination forum, the Group for Democracy Aid (GAD), to increase their leverage vis-à-vis other donors, the UN peace operation (UNOMOZ), its humanitarian arm (UNOHAC), as well as the government and Renamo.

Why were donors so strongly committed to making the peace process work? Additionally, why were they able to overcome the familiar obstacles to coordination of aid? First, donors clearly understood the importance of a successful peace process for the achievement of their own goals for the country, whether their priorities were macroeconomic stability in a market-oriented economy, food security, ecologically sustainable development, or improvements in health or education for the population.

Second, donors were keenly aware of the consequences of failure, due to the recent collapse of the peace process in Angola. Howard (Reference Howard2008: 194) points out that, in Mozambique, ‘the formulation of the mandate and its implementation were conducted under the fear of failure recently experienced in Angola’. There, a small UN peace mission with a weak mandate had been unable to oversee a successful transition. Elections in September 1992, intended to be the culmination of the peace process, instead led to renewed war between the government and UNITA (Anstee Reference Anstee and Crocker1999; Howard Reference Howard2008; Jett Reference Jett1999). Some of those hired to provide technical assistance to Mozambique's electoral process had worked in Angola during this time, and they were eager to impart the lessons learned from that experience to donors (Turner Reference Turner and Mazula1995: 644).

Donors, hungry for a success story, were ready to listen. The Angolan debacle convinced the international community of the need for a stronger and better-funded UN peace mission in Mozambique. At the same time, this situation also made the donors more sensitive to the limitations of UN peacekeeping. In Mozambique, the UN was slow to deploy, and there were coordination problems right from the start between UNOMOZ, UNDP and other UN agencies on the ground. UNOHAC, the entity created by UNOMOZ to coordinate aid, was often hamstrung by its own bureaucracy, and was staffed by personnel with less experience than many who worked in the UN's specialised agencies in Mozambique or in relief organisations there (Ball & Barnes Reference Ball, Barnes, Forman and Patrick2000).

Donors, then, had the motivation, the knowledge, and the experience necessary for the kind of donor coordination that would make peace conditionality possible. Yet there was little reason for UNOMOZ to cooperate closely with bilateral donors, and such cooperation would be critical. A full explanation of the successful peace operation must include the leadership style and personal predispositions of the head of UNOMOZ, Special Representative of the Secretary General (SRSG) in Mozambique, Aldo Ajello. Ajello, an Italian politician with good contacts in Italy and New York, communicated regularly with the ambassadors of major donor countries, preferring direct contact and confrontation, rather than suave diplomatic manoeuvres. He prioritised transparency and consultation in his relationships with domestic political actors as well as ambassadors of the major donor countries. In fact, as one contemporary noted, Ajello was ‘willing to bend the rules when it was critical to the success of the mission’ (Jett Reference Jett1999: 77). His willingness to include donors as partners in the peace implementation process was essential.

CONDITIONALITY AND SUPPORT FOR ELECTIONS

It is in supporting electoral processes that donor coordination and conditionality were most effective. Elections were to be the capstone of a successful peace process in Mozambique, and they were funded almost entirely by donors. The Government of Mozambique met US$5 million of the $60 million price tag for elections, with the rest provided by the donors (Howard Reference Howard2008; Jett Reference Jett1999: 110). Donor coordination around the electoral process arose as a direct outgrowth of dissatisfaction among a group of donors with the logistical and financial arrangements for transitional elections. The key donor coordinating mechanism for elections was the GAD, a group of major bilateral and multilateral donors formed in 1992 to ‘monitor donor support for various democratic initiatives that were planned as a result of the Rome peace negotiations, then in progress’ (Turner Reference Turner and Mazula1995: 651). The GAD was coordinated by the UNDP Resident Representative and its principals included the USA, EU, World Bank, Norway, Sweden and Denmark. These donors already had some experience working together in Mozambique, as members of the Emergency Programme donor group, which had liaised on the provision of emergency relief to Mozambique for several years before the peace agreement.

The GAD facilitated the work of bilateral donors who were committed to making elections work in Mozambique in several ways. First, it enabled the exchange and circulation of information among different donors. This gave donors regular opportunities to discuss linkages, sequencing, and priorities within a complex, multifaceted peace operation. Like other coordinating forums, it also helped prevent or dispel misinformation about the peace process. As one observer points out, the regular channels of communication between donors ‘proved critical in an environment characterized by rampant rumors. Because the donors were so well informed, they were able to reassure a suspicious and insecure Renamo leadership and to reduce tensions between Renamo and the government’ (Ball & Barnes Reference Ball, Barnes, Forman and Patrick2000: 186).

Second, the regular meetings, bolstered by reports from UNDP election experts who had recently arrived from Angola, created a shared understanding of the importance of these elections for the peace process as a whole. Close cooperation among external actors in Maputo helped the ambassadors of donor countries to secure the necessary resources for elections. GAD was chaired by the resident representative of the UNDP, which was given the important role of providing technical assistance to Mozambique's electoral authorities, working alongside the National Elections Commission (CNE) and the Technical Secretariat for Electoral Administration (STAE) as advisers. This provided UNDP with ‘a privileged channel of information about technical and financial aspects of the electoral process, information which became indispensable for keeping donors informed of the progress being made, as well as the needs that existed’ (de Mul Reference de Mul and Mazula1995: 639)

It was important that Aldo Ajello, as SRSG and head of UNOMOZ, himself valued and participated actively in GAD, as well as consulting regularly with the ambassadors of countries represented on the Supervisory and Monitoring Commission (USA, UK, France, Italy and Portugal), the representatives of the permanent members of the Security Council, and the Like-Minded Group (Ajello Reference Ajello and Crocker1999: 628). ‘In this way’, Ajello observes, ‘all the key international players were briefed extensively and views were coordinated in advance of reporting. Hence, each representative was able to send a simultaneous report to his or her capital, reflecting a comprehensive and common political analysis of the situation. This usually generated a common reaction’ (ibid.). In addition to creating the image of a united front to the warring parties, GAD and other ad hoc coordinating groups helped Ajello and the leaders of other bureaucratic organisations – including donors who had to report to their own governments – to surpass their own limitations by providing them with extra-organisational allies able to frame and help implement creative solutions not foreseen in New York, Washington, Rome or London, nor contained in the UN mandate.

Finally, and certainly not least important, GAD served to hold its own members to their pledges of support. Michael Turner's (Reference Turner and Mazula1995: 650) first-hand account provides an example. (Turner was the USAID Democracy and Governance adviser during this period and a participant in GAD meetings.) He notes that he ‘clearly recalls an experience when the internal procedures of USAID were slowing down initial support promised for the registration campaign and having been publicly “reminded” [in a GAD meeting] that the contribution promised by the government of the United States to rent small planes was vital for the successful start of the registration campaign. Consequently, the planned visit of the author to the province of Zambezia was delayed in order to complete the necessary documents.’

Over time, the regular meetings of this group began to forge a sense of institutional identity. The members of GAD established themselves, and asserted their identity, as the leading donors in the area of electoral assistance in Mozambique. As such, they gained legitimacy and sometimes leverage with respect to both other external actors, such as UNOMOZ, and the government and Renamo (Turner Reference Turner and Mazula1995).

One important example is the group's successful push to expand the agenda of an important donor conference to include electoral processes. This June 1993 conference was envisaged as an opportunity for donors to pledge resources specifically for the peace process (unlike the Consultative Group meetings, which continued to deal primarily with emergency relief, though this, of course, also touched on elements of the peace process). Despite the fact that the conference was specifically designed to procure funding for the peace process, its convener, the UN DHA in New York, had made no plans to discuss elections at the conference. As Turner (Reference Turner and Mazula1995: 658) argued, donors who had been working in Mozambique – as opposed to conference planners in New York – knew that ‘all the components [of the peace process], including the electoral process, were interrelated and needed to be seen and analyzed as a whole’.

UNOMOZ and the principals of GAD, along with a few other bilaterals, organised a series of parallel sessions before and after the first session of the official conference, and issued a declaration emphasising the importance of the elections to the process and their commitment to funding them. This declaration was subsequently accepted by the UN in New York, and electoral processes were henceforth viewed as an integral part of the peace process. This was important in facilitating further cooperation between donors supporting electoral processes, and in justifying those donors' requests to their own governments for adequate funding for these activities. It also reinforced the standing of this ad hoc donor group as a key player in the peace implementation process, and strengthened its members' confidence in and commitment to this coordination mechanism (Turner Reference Turner and Mazula1995: 659).

By providing information and facilitating cooperation and coordination among donors already committed to a successful peace process, GAD made possible the effective application of successive peace conditionalities by donors. One early example is provided by the Multiparty Conference. The government had organised this conference to draft electoral legislation, convening all opposition parties and inviting donors to observe the process. In a September 1993 memorandum addressed to ‘all Mozambican political parties’, GAD expressed its concern at the collapse of the Multiparty Conference and urged opposition parties to trust in the good faith of the government and to be mindful of the time frame for elections. ‘We stand ready to do our part, but we call upon the Government, Renamo, the emerging parties and the Mozambican people to assume the historic challenge and responsibility they face’ (GAD 1993). They reminded the parties that ‘our commitment to fund elections cannot be open-ended, particularly in view of other demands for scarce resources’, and advanced a proposal for the party composition of the National Elections Commission, which had become a major obstacle.

Conditionality in this case was based not on acceptance of the donors' specific proposals for particular points of electoral legislation, but on progress and cooperation among parties. By insisting that all actors cooperate to craft the rules, donors applied means-oriented as well as goal-oriented conditionality. It is also worth noting that conditionality came as part of a coordinated effort by all donors who were interested in funding the electoral process – here again, donors wanted to ensure that their resources were invested effectively.

The same concerns were manifest in donor support for political parties. Mozambique was one of the first cases in which donors provided aid expressly for the development of viable political parties. With the support of the country's major donors, the UN created two trust funds for the purpose: one to support all registered political parties (seventeen parties received US$150,000 each), and one to support the transformation of Renamo into a political party. The UN Trust Fund mechanism was created to alleviate the misgivings and restrictions of many donor countries with respect to funding political parties directly (Manning Reference Manning2002). However, the initiative was spearheaded and funded by bilateral donors.

Their hesitation about this prospect was palpable, but donors understood that the existence of viable opposition parties was key to the success of the peace process. As a cable from the Heads of Mission in Maputo to the EPS Africa Working Group (EC 1993) put it, ‘the most pressing issue concerning the elections is the proper involvement of Renamo – and the other emerging political parties – in the process’. The Renamo trust fund was a particularly challenging proposition for members of the Like-Minded Group, who had forged strong bonds with Frelimo over the years. Nevertheless, thirteen countries, plus the European Commission, contributed to that fund. Italy made by far the largest contribution, over US$11 million, followed by the USA and the EU. Denmark, the Netherlands, Sweden and Norway contributed a combined total of over $1.96 million to Renamo's trust fund (twice the US contribution), with Netherlands the fourth largest single contributor (UNDP 1995). The contributions of the Like-Minded donors, who had strong and longstanding ties with the Mozambican government that were based on shared ideology from independence on, make clear how important these donors believed it was to secure Renamo's full participation in the political process, and the lengths to which they were willing to go to underwrite success.

This concerted donor support for Renamo – and more specifically for its leader Afonso Dhlakama – permitted the effective use of conditionality vis-à-vis that organisation's participation in the peace process. Renamo's trust fund was premised on the transformation of Renamo into a system-supporting political party. This bargain required Renamo to identify, recruit and train personnel to fill its places on the many peace commissions and on provincial electoral lists. It also required the creation and staffing of a national office, the mounting of a nationwide political campaign, and, of course, the demilitarisation of its structures through demobilisation, disarmament, and redeployment or social reintegration of its personnel. Though by his own account Ajello, who was responsible for administering the trust fund, gave Dhlakama considerable leeway in the use of his monthly stipend (about US$300,000 at its peak), this was money with many strings attached (Ajello Reference Ajello and Crocker1999).

The intensive and on-going involvement by the major donors in Renamo's incorporation into Mozambican political life gave these donors critical leverage on 27 October 1994, the eve of the transitional elections. Convinced that Frelimo, in collusion with SADC states, had conspired to ensure his political defeat, Renamo president Dhlakama announced his party's withdrawal from the electoral process. It was intervention by individual donors, with the support of Ajello, that convinced Dhlakama to abandon his boycott of the elections. The American ambassador at the time recalled how ‘that evening I met alone with Dhlakama for over an hour and repeated that message [that elections would be held as scheduled with or without him]. I also assured him that any claims of electoral fraud would be investigated seriously. Shortly after we met, and I was virtually the only visitor he received that evening, he instructed his supporters to participate in the elections’ (Jett Reference Jett1999: 111). Ajello (Reference Ajello and Crocker1999: 635) recalls how ‘the entire diplomatic corps in Maputo worked with me to mobilize the maximum number of heads of state and government to send an unequivocal message to Dhlakama that the international community did not understand his decision and that he was losing all the credibility he had accumulated over the past two years’. Dhlakama's boycott ended only after he received written assurances from the international community that any complaints of fraud would be thoroughly investigated and the elections invalidated if they were indeed found to be fraudulent.

The fact that it was assurances of international intervention to investigate and deal with any credible accusations of fraud that brought Renamo back into the process is telling. The links between Renamo and the international community were stronger than links between Renamo and the multiparty peace commissions (including the CNE) on which it was represented. Renamo re-entered the elections because it had received guarantees from the international community. In Dhlakama's words (Manning Reference Manning1997: 243), ‘I have always believed that there will be a positive response whenever one presents his legitimate reasons and succeeds in putting them across to the international community. Therefore, when we saw that the CNE was not reacting, and when people in Mozambique were showing contempt for our complaints, we had to make a decision. Ours was a carefully made decision because it permitted the international community to take into account a number of aspects that could hinder the electoral process.’

The international community was seen by Renamo, and clearly saw itself, as a credible guarantor of the process. It was also acknowledged as such by the government, which routinely chafed at the infringement on its sovereignty posed by the large UN mission, and the interventionist approach of its leader and his partners in the donor community. After the coordinated interventions of Ajello and donors in the demobilisation and reintegration processes, and in the Renamo boycott, and given donor financial support for and monitoring of elections, it was clearly not UNOMOZ alone that served as the credible guarantor, but UNOMOZ as assisted, supplemented, and sometimes circumvented, by a determined and capable group of donors.

Peace conditionality, ‘the use of aid as a lever to persuade conflicting parties to make peace, to implement peace accords, and to consolidate peace’, was employed in Mozambique on two levels (Frerks Reference Frerks2006: 1). First, a large infusion of aid was committed by the country's major donors upon signing of the peace agreement in Rome in 1992. This commitment of aid was based on the expectation that the government of Mozambique and Renamo would move forward to implement the commitments each had made in Rome.

However, as the peace process unfolded, donors involved themselves in the details of the implementation process in ways that enabled them to exercise an implicit conditionality over specific measures. The broad quid pro quo of money for progress in the peace process developed into the use of more narrowly targeted peace conditionality to overcome specific obstacles over the course of the implementation process. UNOMOZ as an institution was constrained both by its mandate and by resource constraints, and strictly limited its involvement or the sharing of its equipment and other resources for other parts of the electoral process for which it was not specifically tasked (its priorities were electoral observation and administration) (Ball & Barnes Reference Ball, Barnes, Forman and Patrick2000). Bilateral donors stepped in at many points during the implementation and electoral processes to offer support where UNOMOZ could not. These bilateral actors not only achieved an unaccustomed degree of coordination among themselves with respect to important goals and activities in support of the political transition; they also succeeded in limiting the government's ability to ‘divide and conquer’ the donor community by allowing donors to speak with one voice on the most important issues of implementation. The fact that Ajello, as SRSG, so visibly supported these efforts by bilateral donors was crucial.

In addition, donors sometimes conditioned their own donations on the participation of a critical mass of other donors, to ensure that their own contributions would not go towards a fatally underfunded project. In these ways, donors not only selectively exerted leverage over leaders of the warring parties, but also over UNOMOZ and one another, shaping and reshaping key aspects of the process along the way.

Thus, the specific peace conditionalities that helped produce a successful transition from war to peace in Mozambique were not planned in advance of the peace operation, but developed by donors in response to specific challenges that arose over the course of implementing the peace agreement. When problems arose in the implementation of the agreement, donors stepped in, singly or in groups, to provide good offices, inject additional resources, and remind both sides of their commitments.

Donors' history of engagement in Mozambique provided the necessary backdrop for the application of micro-level conditionality in support of the peace process. By the time of the peace process, the major donors were known quantities to the government and Renamo. The Like-Minded donors had been intimately involved with supporting the state since independence. The USA, Italy, the World Bank and other donors provided significant support later on, but were established actors well before the war ended. This is not to say that relations between these donors and the government or Renamo were uniformly warm. Nevertheless, major donors had established a track record with the government by the time of the peace agreement, and their priorities and proclivities were clear. This was a very different government–donor relationship than the conditional sets of relationships that often emerged elsewhere in the 1990s.

Extended experience in the country gave donors the local knowledge necessary to sense what kinds of positive and negative inducements were most likely to work. Conditionality could be fine-tuned through consideration of the context. Moreover, donors had significant experience working with one another in Mozambique in the context of the humanitarian emergency during the civil war. This experience, their shared understanding of the political, social and economic contexts, and the absence of complicating factors like lucrative natural resources or strategic political or economic importance, made donors more willing and able to cooperate with one another.

Mozambique's experience supports the arguments of scholars who find that peace conditionality is likely to work best where conditions imposed are specific, flexible and well matched to the problem, where donors share common goals, are able to create coordination mechanisms, and where short-term conditions are part of a longer-term relationship and set of goals. Thus, while donors used aid to overcome specific sticking points, these punctual interventions were clearly part of a set of longer-term goals that donors would continue to support for many years after the formal transition. These included macroeconomic stabilisation, not least to provide a favourable investment climate for foreign investors, strengthening the management of public finance and the overall capacity of the Mozambican state to provide basic services, and democratic governance. In the years after the transitional general elections in 1994, donors have continued to support Renamo financially, to provide support for election implementation and monitoring, and to play an active part in monitoring and conditioning their aid, on the requirement that both the government and Renamo remain committed to maintaining peace and moving forward with democratisation.

The Mozambican case reinforces many of the findings from the literature on the role of third-party guarantors in securing negotiated peace settlements. Specifically, it provides insight into the complex interactions that underpin effective leverage for outside actors during the peace implementation process. For example, a number of scholars find that belligerents in both interstate and civil wars are influenced by ‘audience costs’ (Doyle & Sambanis Reference Doyle and Sambanis2006; Fortna Reference Fortna2004; Walter Reference Walter1997). Fortna (Reference Fortna2004: 213) argues that ‘the international community has a strong effect on belligerents’ decisions about war and peace. States worry about international audience costs.' Audience costs are the opportunity costs of breaching the agreement, and are higher when international actors are able to monitor implementation closely and respond with rewards or punishment as appropriate. The effectiveness of audience costs is often diminished because UN peace operations tend to be multifaceted, with responsibility for overseeing the ceasefire, elections, human rights, and more. Fortna (Reference Fortna2004: 205) finds in her study of interstate war that ‘the effects of international audience costs are often limited, either by the UN's desire to maintain neutrality or by great powers turning a blind eye for strategic reasons’.

The same often holds true in post-civil war peacekeeping. Typically, leaders of peace operations are forced to prioritise between, say, timely elections or compliance on the ceasefire, and accountability on human rights. Efforts to take a firm stand on one particular issue may face opposition by powerful members of the Security Council, or compromise the mission's ability to carry out another dimension of its mandate. The literature suggests that while third parties can affect the behaviour of belligerents in positive ways, the terms of their involvement tend to limit their effectiveness in this regard.

In Mozambique, this problem was avoided because of the intensive, direct involvement of bilateral donors in areas that were technically covered by the UN mandate, and because of the active cooperation of the SRSG with bilateral donors. Nevertheless, the structure and multifaceted nature of the mission gave UNOMOZ itself little effective leverage in some of these areas. The UN had formal authority to oversee demobilisation and elections; bilateral donors had the resources, knowledge and practical capacity to make it work. Bilateral donors supplemented UN funding for elections, party transformation and DDR, and played a direct role in monitoring these aspects of the agreement. Even if the UN itself was forced to turn a blind eye to violations of agreements on one or more of these dimensions in exchange for success on another, or to ignore government transgressions in order to continue to operate, bilateral donors – working with the blessing of the SRSG – had the ability to impose direct conditionality, and they did so.

When the parties engaged in excessive foot-dragging on key measures of electoral law during the Multiparty Conference, donors withdrew funding for the conference facilities. When Renamo pulled out of the elections on the eve of voting, donors promised to investigate Renamo's concerns. When donors feared demobilised soldiers would become a source of instability in the countryside after their severance pay ran out, they designed and funded a programme to extend their benefits (Manning Reference Manning2002). None of this would have been possible for the UN peace mission alone, whose multifaceted mandate made it incapable of taking calculated risks in one area, out of fear that this might compromise its role in others. The direct involvement of bilateral actors allowed the international community as a whole to fine-tune its role as a credible guarantor of the peace agreement.

This involvement did not come without costs. The major donors and the SRSG treated the peace process as a one-time intervention in which the ends justified the means, and in some respects placed a firewall between the peace process and longer-term processes of political and economic development and post-war reconstruction in the country. While this perspective made donors more flexible and innovative in responding to the challenges of the peace process, it also involved trade-offs. The peace process clearly reinforced the donor community's position as a rival to government sovereignty, with implications for Mozambique's longer-term development strategies and prospects (Abrahamsson Reference Abrahamsson1995; Hanlon Reference Hanlon2008).

Moreover, there was no attention in Mozambique to transitional justice, or to the longer-term effects of massive external support for Renamo. What would have been the effect on the peace process if the international community had decided instead to criminalise Renamo, as with the RUF in Sierra Leone or the Lord's Resistance Army in northern Uganda? How important was the decision to invest instead in Renamo's legitimation, and with what longer term impact on the country's politics?

The comparative literature on post-conflict peacebuilding has focused on the UN as the major independent variable, while controlling for domestic political, social and economic context, and for the nature of the war. The Mozambican case points to the need for more fine-grained investigation into the diversity of the actors and micro-level approaches that make up complex, multidimensional peace operations, whose success or failure is too often assumed to hinge on the UN mission itself.