The Status of Arms Flows in Mozambique

"People are reluctant to disarm when they are not sure of the nature of the world in which they would live without arms"
(Nye, 1993:18)

The end of the civil war introduced a new phenomenon into Mozambique, with serious consequences for its neighbouring states. With the advent of peace, large numbers of surplus weapons flooded the country. Given the prevailing deprivation of much of the population and the decline of the state and its consequent lack of control, these weapons are being used in criminal, illegal and violent activities. The control of small arms has, in consequence, become a prerequisite for regional stability and development.

Studies conducted in the area of arms control have placed emphasis on the need for control of arms transfers between states or, at most, from manufacturers to subnational organised groups. In Mozambique, such a state-centric view could have been effective either during the liberation war against Portugal or the recently ended civil war. Neither the global nor the regional contexts, however, allowed such control. The result was that an inestimable number of weapons entered the country both legally, through government procurement, and illegally, through covert supplies to guerrillas.

The nature of the wars and the weakness of the warring parties’ logistical control mechanisms led to arms diffusion down to the level of the individual citizen. Subsequent arrangements for the termination of war have, in turn, failed to retrieve these arms from Mozambican society. This diffusion of weapons began when the Lusaka Accord of 1974, marking the end of the liberation war, failed to make any provisions for disarmament. The result was that the weapons used during the war remained in the country and have continued to spread among the freedom fighters, as well as the general population who have been issued with guns for self-defence.

The civil war only exacerbated weapons diffusion. During that time not only did the parties acquire arms in excess, but they also distributed them naively to militia groups and the general population. This was the situation confronting the drafters of the General Peace Agreement (GPA) and the subsequent United Nations Operation in Mozambique (ONUMOZ), the agency responsible for implementing disarmament and arms control. From the beginning, however, the UN process failed to define the ONUMOZ mandate correctly. In fact, disarmament was defined in traditional terms in which only politically organised groups are armed. But, in Mozambique, the situation had gone beyond that. By restricting the mandate to military institutions, therefore, a gap was created, leaving a great number of arms outside of the disarmament process. In addition, even the so-called ‘one-man/one-weapon’ disarmament principle was not effectively carried out. As a result, when in 1994 multiparty elections were held, the country was flooded with weapons, the possession continuum extending from the state level through to the ordinary, individual citizen. In the deprivation that prevailed, the economic difficulties, in particular, armed violence quickly became a means of livelihood and profit.

Bearing all these factors in mind, the present study examines small arms flows in Mozambique. Since it would be futile to attempt to make accurate quantitative estimates of such arms flows, it identifies, instead, the actual patterns of flows in the country in order to make certain policy recommendations. There are two main emphases. First, the historical background of the spread of arms throughout the country is looked at and consideration given, on the one hand, to how arms entered the country and, on the other, to how they got out of control. Thus it aims to give a rough picture of the numbers of arms that were brought into Mozambique, as well as to show how they were used during the two conflicts. Secondly, taking into account this historical data, it looks at the present situation in order to explain the patterns of arms flows. It is suggested that there are presently, in Mozambique, three categories of arms possessors, among whom arms can circulate quite freely. Finally, it notes that the most important sources of arms in Mozambique are the caches deliberately held by the warring parties and leakages from the official armouries, concluding with a policy recommendation in terms of which the former warring parties should take political responsibility for the caches, helping the state institutions to destroy them, and most importantly, the state is advised to police the police before the latter polices the public.

Small Arms Flows in Mozambique, 1964-1994

Historians would be accurate in arguing that the first firearm to enter Mozambique was by way of the Portuguese crown, as late as the 15th century. Even in the anti-colonial resistance undertaken by the indigenous population, during the periods of penetration and effective occupation, traditional spearheads and arrows were still used.

Two factors, however, seem to have prevented a strong Portuguese military deployment in Mozambique before the liberation struggle. First, the backwardness of the war methods and equipment of the indigenous population1 meant that little military effort was required for conquest. Secondly, the Berlin Conference plan avoided a major clash between colonial contenders, particularly between the Portuguese and the British over Mozambican territory.

Even if Isaacman’s claim that the 1926 coup makers were "committed to using the nation’s military resources to guarantee Portuguese hegemony" is taken into account, it remains true that "... [r]ather than injecting the colonies with skilled workers and additional capital, the majority of Portuguese arriving in Mozambique were peasants, whom the government viewed as a drain on the metropole’s economy".2 Mozambique, therefore, was not to witness a massive influx of arms before the national liberation struggle of the early 1960s. Prior to that, the Portuguese military presence in Mozambique had been restricted to broad maintenance of public order.

The Impact of the Liberation War 1964-1974/5

Unlike the other colonial powers in Africa, Portugal failed to grant its colonies a peaceful independence. Hence, following the normal trend of African nationalism of the 1960s, the Mozambican Liberation Front (FRELIMO) launched the struggle for liberation in 1964.3 Initially, FRELIMO’s activities were concentrated in the northern part of the country, since it had few trained men and few weapons.

Following an upsurge of more violent war activities, in 1970,4 the Portuguese general, Kaulza de Ariaca, was given a force of 35 000 men and 15 tons of arms with which he launched the notorious Gordian Knot.5 Meanwhile, FRELIMO were reported to have some 10 000 armed guerrillas trained in Algeria, Tanzania, Cuba, China and Yugoslavia and another 10 000 men without arms.6 If these numbers are taken as reliable, then one can estimate that in 1971 Mozambique had about 45 000 small arms,7 excluding stocks belonging to the security forces.

The more FRELIMO intensified the war, the larger the Portuguese military presence became and the more numerous the weapons that entered the country. At the beginning of 1974 there were reported to be 60 000 Portuguese troops stationed in Mozambique. Three months later, a further 10 000 were sent in from Angola.8 On the FRELIMO side, it was difficult to estimate how many armed men there were. At the time of independence, FRELIMO still claimed some 10 000 troops, equipped predominantly with Soviet arms, including 60 mm and 82 mm mortars, 122 mm rocket-launchers, SA-7 surface-to-surface missiles, 14.5 mm anti-aircraft guns and some Mig fighters.9

Therefore, the first major flow of arms to Mozambique took place between 1964 and 1974 to 1975, fuelled by the liberation struggle. At the height of the war, Portugal had an armed force of about 70 000 men. As Berman points out, the FRELIMO 10 000-man force "... was smaller than the number of people that had taken up arms as part of the militia during the liberation war."10 Of the total number of arms entering Mozambique, the percentage of light weapons was unquestionably high, the main reason being that FRELIMO was a guerrilla movement using predominantly small arms and light weapons.11 If one takes into account that the nature and capability of the enemy are among the factors to determine the type of arms imported into a country, that may further have influenced the type of weapons procured by the Portuguese.

However, as far as the impact of the liberation war on small arms proliferation is concerned, three further points should be made. First, during the liberation war, many weapons entered Mozambique, though the actual number is unknown. These were for use by FRELIMO guerrilla fighters and the Portuguese colonial army, the East-West geopolitical divide facilitating arms dispersal to both parties. In fact, even without hard currency to pay for weapons, FRELIMO was able to get concessionary deliveries from the Soviet Union, China and some East European countries. Meanwhile, within the NATO arrangement, Portugal benefited from the US policy of ‘benign neglect’12, as well as from general western verbal condemnation of colonialism.13

Secondly, both FRELIMO and the Portuguese army had very poor arms-control mechanisms. FRELIMO, as a popular guerrilla movement, had involved the civil population in the handling of arms, distributing them to civilians in the liberated zones both for self-defence and as part of the strategy of lending a popular nature to the war. Unfortunately, this had the negative effect of helping to "legitimise armed action and spread of weapons throughout the country".14

In a similar way, the Portuguese colonial authorities distributed weapons to white settlers. According to African Contemporary Record reports, in Beira, following a three-days’ demonstration against the failure of the army to provide security, large numbers of weapons were distributed by the authorities. This became such common practice that between May and June 1974, a total of 5 000 licences were issued in favour of settlers.15

Thirdly, the 7 September 1974 Lusaka Accord made no provision regarding demobilisation and disarmament. It provided only for the transference of power to FRELIMO. As far as military issues were concerned, the guerrilla organisation was to take over gradually, through a process of transformation into a national army; while the Portuguese army was to be dismantled. To be sure, "... in the first month of the new [transitional] government’s office, some FRELIMO troops arrived in the capital and the Portuguese Government announced it was withdrawing 12,000 troops from the country".16

There was no mechanism for retrieving the arms that the Portuguese authorities had distributed, nor was there a clear political directive concerning what should be done with the weapons in the aftermath of the war. In addition, the locally conscripted Mozambicans who served in the Portuguese army were sent home without any plans for their social or economic reintegration into society.17 To make matters worse, they were shunned by FRELIMO leadership as the agents of colonialism.

Set against this, FRELIMO’s concept of demobilisation in the post-Lusaka period has never been clear. The so-called ‘liberation war veterans’ were never officially demobilised, although some had to leave the barracks. In general, they were not encouraged to return to their places of origin, but were quartered in rural or suburban districts, most of them retaining their weapons. In line with this, an undocumented reintegration support scheme was undertaken to allow their socio-economic integration. The privileged position of these pseudo-demobilised veterans can be demonstrated by the establishment within the cabinet of a state secretariat for war veterans. This treatment seems to indicate that FRELIMO wanted them to remain loyal and easily able to be remobilised, not necessarily for internal matters, but in the event of external aggression. In practical terms, however, this meant FRELIMO’s loss of the available arms in Mozambique. In hypothetical terms, it might be argued that, even if the war with RENAMO had not occurred, violent crime would inevitably have followed.

The Impact of the Civil War 1976-1994

Within ten months, from September 1974 to June 1975, FRELIMO became the government of Mozambique and the Frente de Libertação de Moçambique (FPLM) became the Mozambican National Armed Forces (FAM). Initially, Machel’s military concern was to modernise and professionalise his forces in order to transform them from a guerrilla group to a regular army.

However, in neighbouring Rhodesia, the Ian Smith regime used a new impetus of the old diplomatic and military support for Robert Mugabe’s ZANLA (Zimbabwean African National Liberation Army), coupled with the 1976 trade embargo to and from Rhodesia, to justify both raids into Mozambican territory and the lending of support to the emerging Resistência Nacional Moçambicana (National Resistance of Mozambique (RENAMO)).18 Machel was forced, as a result, to increase the army from the alleged 1975 level of 10 000 men to 21 200 troops within three years. Of course, such massive recruitment jeopardised Machel’s dream of professionalisation and, from 1976 onwards, defence and security were made top priorities in the budgetary law.19

When Rhodesia became Zimbabwe, RENAMO moved to South Africa, where it came of age. The reasons for RENAMO’s transfer to South Africa in 1980 are widely known and similar to those which had led Smith to foster the movement in the first place. Following this move, Machel’s government further increased the defence and security share of the government budget to estimated levels of 40 to 50 per cent.20 This figure did not include traditional Soviet military assistance given on credit. Indeed, Russian equipment and training support were reported to have continued in 1977 on a larger scale than before. Western intelligence services claimed at the time that some 6 000 tons of military supplies had been shipped to Nacala Port in February 1977 alone.21

These developments characterised the government’s response to the events taking place throughout a full, albeit short, transition period from a liberation war (ending in 1974 with the Lusaka cease-fire agreement outlined above) to a civil war (which quickly acquired a life of its own, because of regional dynamics). The richness and complexity of this period stem from the regional significance of Mozambique’s independence and its ideological and political orientation.

The achievements of 1974 and 1975 were never translated into peace, and hence arms flows into the territory were not halted. A combination of factors at different levels resulted in one of the most violent civil wars in the region. On the domestic level, the exclusive state-party system had created a disenchanted elite; the FRELIMO radical regional policies challenged the prevailing status quo; and, finally, the second Cold War of the 1980s created a global context for weapons transfers to the Third World22 and to Southern Africa, in particular. By 1980, both RENAMO and FRELIMO had received large amounts of weaponry from outside suppliers. In order to have a clearer picture regarding the numbers of weapons concerned it is necessary to scan briefly the historical evolution of both forces.

According to Berman, RENAMO started as a group numbering a few dozen. In Rhodesia it grew to some 1 000 to 2 000 men, but during the first half of 1980, as a result of Zimbabwean independence, the group was reported to have been reduced to a force numbering between 250 and 500. Following its transference to South Africa, however, it had grown rapidly again and in 1982 it was believed to be between 3 000 and 8 000 men strong. Two years later, in 1984, when there was a South African sponsored attempt at a negotiated solution, RENAMO was believed to have some 16 000 men. Despite the tempting amnesty launched between 1988 and 1989 by the government in the hope of enticing the guerrillas to abandon the fight, RENAMO had, by the end of the decade, more than 20 000 fighters.

As far as military supply is concerned, historically RENAMO received weapons from white minority regimes in Southern Africa and from both Portuguese and other Western business interests, as well as considerable numbers of captured infantry weapons. South Africa is reported to have supplied significant quantities of Soviet AK-47s captured in Angola and Namibia. RENAMO was supported also by pro-Christian and anti-Communist organisations in the West. The common feature of all these supplies was that they were predominantly light weapons suited to RENAMO’s guerrilla-movement nature.

Logistically, these weapons were distributed to anyone who could fight, including children. Recruitment was generally through kidnapping or conscription, meaning that there was no guarantee of political and ideological commitment.23 As a guerrilla movement, popular support, or at least popular involvement, was vital. Weapons were entrusted to civilians for concealment. This fact not only prevented RENAMO from controlling its arms effectively, but also resulted in these being scattered to every corner of the country.

Until 1980, because the FRELIMO government believed that its most serious security concerns were South Africa and Rhodesia, RENAMO was not taken into account. In preparation for a conventional attack from the two neighbouring giants, emphasis of security-related efforts was placed upon the modernisation and professionalisation of the army. However, from 1980 onwards, increasing military pressure from RENAMO throughout the country led the government to undertake fundamental reorganisation of the army and a massive recruitment drive.

In 1980 three regional commands were created. This meant further mobilisation and armament. In 1982, when about 8 000 RENAMO guerrillas were scattered all over the country, the government created ten provincial commands which were assigned the task of recruiting soldiers to serve at district and local levels, as they considered necessary. It seems that from then on nobody knew how many men the FAM had, though rough estimates reported that, by 1990, Mozambique had around 145 000 people on its military payroll.24 These figures do not include foreign troops assisting the government.25

This was the start of the protracted process of militarisation of Mozambican society. Arms were spread throughout the country, reaching people of all ages, and involving even the family unit. The following figures give some idea of the situation:

in Sofala, a rural province, arms were carried by 40 per cent of the adult population;26



the town of Beira had approximately 30,000 armed militia;27



in 1995 an estimated six million AK-47s were circulating in Mozambique and the government had distributed some 1,5 million assault rifles to the civilian population for self-defence during the civil war;28 and,



there were about 155 600 armed men in Mozambique by the time of the ONUMOZ exercise.29

Although these estimates may be exaggerated, it is an incontrovertible fact that the government had, on the one hand, a large, uncontrollable, disloyal army. On the other, unable to provide its people with security, it had distributed an unknown number of weapons to militia groups and civilians alike, for the protection of cities, suburbs, villages, districts, state owned-companies and other areas.

The distressing fact was — and remains so today — not that such quantities of weaponry were being distributed, but that there was — and still is — no mechanism for control which allowed the military authorities to retrieve these weapons in time of peace. In an interview, a former district administrator30 said that most of the then district administrators had more than four weapons each (including AK-47s and pistols), most of which were not registered by the military authorities. "I gave back two and the others, I do not know where they are", he insisted. In another interview, a former army officer had to admit that the motivation for arms distribution had been purely political:

"The military was not even consulted and, of course, it did not have in mind the need for arms control. From a military point of view the distribution of weapons to [the] civilian population was nonsense. Even to militia groups it should have been more cautious. Because, those arms could - and most of the time they did - end up reinforcing the enemy. But the political leadership demmed it correct!"31

It seemed that, as far as the government was concerned, military consolidation was the only way of ending the war. In such circumstances the soldier, as a human being, was of less importance than his weapon. In fact, "while the government never overcame the obstacles of outfitting the provincial armed forces with uniforms, paying their salaries, or distributing their rations, it was eventually able to provide them with weapons."32

What the consequences of this situation are is beyond calculation. Possibly the most useful way to assess the political effects would be to quote Ayoob’s observation that "... one of the most significant political results of the widespread proliferation of light weapons amongst the warring factions [and amongst the people at large] was[/is] the diminution not just of central authority but also of traditional foundations of order, namely the clan and tribal elders."33 As far as Mozambique is concerned, studies have demonstrated that the recent history of the country has jeopardised the role of the traditional authorities as providers of order, loyalty and social trust.34 This has led to the structuring of a culture of violence where neither order, trust nor loyalty exist; and where the rule of the strong prevails.

These were the hard facts for which the Rome General Peace Agreement and its ONUMOZ implementing agency should have found final solutions.

The process leading to the Rome General Peace Agreement (GPA) on 4 October 1992 has been extensively documented. The driving factor was that the situation had reached such critical levels that neither the warring parties nor the population could sustain the costs any longer.35 Against this background, a permanent constraint, shaping the entire process from the negotiation phase through to the implementation one, was distrust between the parties and, not least, the uncertainty of large masses of the population. The international community, too, experienced doubt as to the capacity and willingness of the warring parties to surrender the military option following the agreement.36

When, following the cease-fire agreement, no serious military violations took place, analysts were taken by surprise. The IISS Strategic Survey 1994-1995 endorses this response: "In Mozambique the most surprising thing about its war was the way it ended. ... [Despite the fact that the atrocities seemed uncontrolled] when the cease-fire was announced, RENAMO stopped the fighting and the war stopped. The gangs had not been out of control." 37

The GPA was a very well organised, structured and detailed document, dealing with almost every important aspect of the peace process in Mozambique. However, as a clear-cut blueprint, it was challenged by the complexity of the process itself, due to the prevailing atmosphere of distrust, inequalities between the parties, general economic hardship and the high levels of militarisation of the society, to name but a few obstacles. This fact was to be acknowledged by Aldo Ajello, the UN Secretary-General’s Special Representative to Mozambique, in a very apt analogy: "I had a very tall man and a very short/small cover. I had to change the cover or amputate the man. I opted to change the cover although this may have created problems both in Maputo and New York."38 This meant that the implementation phase was not merely a technical matter. Some of the agreed issues could, for instance, be found unworkable and hence a process of political negotiation would have to take place within the several commissions.

In examining ONUMOZ disarmament achievements one needs to be cautious. The size of ONUMOZ was clearly influenced by the recent experience of UNAVEM II.39 It seemed then that the former would mark the point at which the focus of UN peacekeeping operations shifted from humanitarian relief and electoral processes to military aspects, disarmament, in particular. However, as Ajello so fully articulated, what prevented ONUMOZ from achieving comprehensive disarmament was the prevailing UN philosophical underpinning: "[W]ar is a result of a political decision: the arms are the instrument of war not its cause."40 In New York, the political will to consent to ONUMOZ’s undertaking of disarmament was strong. But the mandate itself largely fell short in its inability to rid itself of the rules of traditional peacekeeping operations. This came to be the major weakness of the disarmament component of the mission. ONUMOZ was expected to be an enforcing, rather than a peacekeeping, force. However, its mandate remained bound by the GPA, signed by the warring parties. As Smith notes, "[a]greements to undergo demobilisation are difficult to achieve" 41 as such agreements depend upon trust.

Thus, although the UN Secretary General had stated clearly that "... it will not be possible in Mozambique to create the conditions for a successful election unless the military situation has been fully brought under control,"42 the link between demobilisation and disarmament did not take into account the gap that had come into being between the military institution and arms availability: arms existed beyond the traditional bounds of the military establishment, both the liberation war and the civil war having been characterised by a high degree of popular involvement in arms handling. But ONUMOZ was confined to undertaking a limited disarmament only. It had no idea — nor did anyone else — of how many weapons were circulating in the country. What was clear was that the 213 951 collected weapons were too insignificant a proportion to register a success (see Table 1). Yet, if disarmament was seen as a top priority of the UN operation in Mozambique at the outset, why, by the end, was it no longer that important? A partial answer is that, within the United Nations, disarmament remains a politicised component of peacekeeping operations. It becomes important when elections are in jeopardy. Once the arms of the two military factions in Mozambique had been ‘politically destroyed’, as a government official claimed,43 the Secretary-General could declare that conditions had been created for a smooth electoral process. That a rise in violent crime was taking place at the time was not taken into consideration: it was beyond the mandate.

These obvious contradictions between the desire for comprehensive disarmament and the unwillingness to act accordingly stem from the dilemma caused by the discrepancy between limited resources and the political mandate as defined by the GPA, on the one hand, and an effective disarmament plan which would have meant omitting to observe the very principle of the consent of the parties, on the other. Indeed, bearing in mind the weapons availability scenario in Mozambique, the latter would have required that ONUMOZ search weapons caches outside the assembly areas, which its mandate did not permit. It was, therefore, easier for the officials not to question the fact that soldiers arrived at assembly areas with less than one weapon on average.

Because the ONUMOZ mandate was concerned with the use of weapons for political purposes, even having noted that there were a great number of arms in the country, it did not consider stepping up the post-demobilisation disarmament process. A request for a supplementary US $52,5 million budget to further disarmament was refused by the UN Security Council.44 Considering the statement that "... to implement a peace, defined in terms of long term stability, is to focus not just on the sources of violence such as social and political issues but also on the material vehicles for violence"45 this appears in direct antithesis to its theoretical stance and thus quite mistaken.

What has been said so far suggests that the disarmament programme undertaken by ONUMOZ in Mozambique should be assessed on two different levels: its failure to elaborate on an all-encompassing mandate appropriate to the control of arms, on the one hand; and the failure to effectively accomplish the goals defined by the actual mandate, on the other. The failure to distinguish the one from the other may have resulted in certain responsible action being demanded from, and certain recommendations being made to, the wrong people or institutions.

Since many anticipated that the ONUMOZ mandate would mark a shift towards a peace force with a focus on enforcement, while its mandate had already been defined within the traditional peacekeeping framework, the gap between expectations for a comprehensive disarmament programme (including dismantling caches outside the assembly areas and collection of arms in the possession of civilians) and the actual more traditional and narrowly-defined disarmament outlined by the parties widened.46 At this level one should blame both the parties and the international community. A realistic assessment of ONUMOZ’s performance would be one that concentrated on the question of what should have been, but failed to be, achieved, within the limitations of the mandate.

Demobilisation constituted the most difficult and dangerous aspect of the peace process in Mozambique. According to the ONUMOZ plan, the soldiers were to be assembled in selected areas and those who were to be demobilised would receive severance payment, subsidies and assistance with transportation to a place of choice. Accompanying this process, there would be the collection, storage and destruction of weapons in the possession of the soldiers. This was the concept of disarmament that prevailed in the GPA in Rome and subsequently guided ONUMOZ activities. Accomplishment of this goal would have meant partial disarmament only. Yet, even this was not achieved. Four main reasons exist for the failure to effect this one-man/one-weapon goal.

First, most of the soldiers had more than one weapon so they could deliver one and keep the others. Of course, in a situation where, "... [e]very time a FAM soldier was transferred to a new post during the war he was issued with a new weapon regardless of whether he still had others or not",47 and given that there was no record keeping, one-man/one-weapon was clearly an insufficient objective for complete disarmament. Related to that was the fact that the weapons brought in by the soldiers were largely of poor quality, which suggests that those were not the weapons most recently used in combat.

Secondly, there was no political will on the part of either party to surrender its weapons completely. It seems that both sides ordered their soldiers to hide some arms. Although they were represented on the cease-fire commission, which was in charge of supervising demobilisation and disarmament, they made no effort to prevent soldiers reporting to assembly areas without weapons. Furthermore, the UN Secretary General regretted the fact that, "... contrary to the rules approved by the cease-fire commission, the United Nations was denied permission to collect and disable weapons in the unassembled locations on several occasions ...".48 Yet, given the leverage they had, ONUMOZ and the international community as a whole could have put greater pressure on the parties to turn in a greater number of weapons.

Thirdly, ten percent of the registered soldiers did not report to assembly areas and thus their weapons were not collected. The main reasons for this were war fatigue, and lack of information concerning the whole process of demobilisation and integration into the new army. Generally, however, few soldiers wanted to join the new army. So having no information about whether the integration would be compulsory or not, and given the experience of conscription and kidnapping into the former armies, they preferred to desert before demobilisation.

Finally, there were the security conditions relating to the storage of collected weapons. The unsafe conditions under which the weapons were kept and the leakage which resulted have been extensively outlined by several analysts.49 It is widely believed that when ONUMOZ left Mozambique a great number of weapons held at government armouries or assembly areas were stolen and found their way back into civilian or criminal possession.50

When, at its departure, ONUMOZ stated that it had collected 189 827 weapons,51 it is clear that the reported success did not relate to comprehensive disarmament of the factions. The post-demobilisation verification process in 744 former military positions and depots, in fact, found substantial numbers of weapons of various types, as well as a small numbers of previously unregistered government and RENAMO military personnel.52 More importantly, although the Secretary-General had indicated that there would be a policy shift with regard to the disarmament commitment, this was never translated into action.

Table 1
Total Numbers of Collected Weapons at the End of the Peace Process

FAM (ONUMOZ + UNilaterally Demobilised. Troops)* 106 799

RENAMO 17 736

Armed, Paramilitary, Private and Irregular Troop 43 491

Post-Demobilisation Disarmament 46 193

Total** 214 219

Source: E Berman, Managing Arms in Peace Processes: Mozambique, UNIDIR, New York & Geneva, 1996: Annex II, p 88

* FAM’s number includes ONUMOZ collected, as well as unilaterally collected, weapons

Small Arms Flow: The Current Situation

A picture of the current flow of arms in Mozambique will now be presented. For ONUMOZ to maintain that it had recovered all weapons from unauthorised hands without its knowing how many weapons there were in Mozambique, or even how many had been held by FRELIMO and RENAMO, was not possible. Even the ex-belligerent parties could be challenged with the fact that they did not know how many arms they possessed. Therefore, the most accurate conclusion seems the one advanced by Oosthuysen: "A combination of factors, such as non-existent record-keeping and the covert nature of supply and transportation deals, has ensured that calculations of the number of weapons in Mozambique can be no more than the educated guess of between one to six million."53

To attempt to arrive at an accurate estimate of small arms in Mozambique would be an exercise in futility, nor is this entirely necessary for developing solutions. What is important, however, is to determine how the arms flow has come about, so that one might ascertain where these arms are most likely to be found. The first distinction to be made is between, on the one hand, the weapons in possession of the FADM and other state security institutions, such as the police and, on the other, those arms in the possession of unlicensed owners and generally used for criminal and violent actions in the society. The former weaponry is made up of those arms which ONUMOZ handed over after the demobilisation process, including the weapons in the state security institutions which were outside the disarmament process. If leakage were not one of the main sources of the arms in unauthorised hands, these weapons would be beyond the scope of the present work.

The real subject of this study is the use of illicit arms for criminal purposes. Given the clandestine nature of these arms, it is hard to categorise them. However, the information issued by the press, police and coming from popular accounts, gives one some idea of what might have motivated people to keep back arms in the post-disarmament period. Hence, in Mozambique, illegal arms seem to exist in three different forms:

caches deliberately retained by the then-warring parties;


caches belonging to individual soldiers, demobilised and in the FADM; and,


arms retained by individual civilians.

An examination of each of these categories gives us the actual picture of small arms flows in Mozambique.

Caches Deliberately Retained by the Parties

As has been made clear, distrust between the parties has been the crucial factor shaping the peace process in Mozambique since its early stages of negotiation. RENAMO was suspicious of FRELIMO’s intentions and wanted to avoid post-settlement marginalisation. Mazula, quoting Matteo Zuppi on the negotiations, states, "... very often things were not going [well] due to several reasons, mainly due to distrust. This distrust was enormous and reciprocal, but it was stronger in the RENAMO side which did not want to succumb to the integration trap ... a strategy with which FRELIMO, during a long period of time, tried to assimilate RENAMO by giving no political value."54 Thus, while RENAMO tried, at the negotiating table, to get as many FRELIMO concessions as possible, it did not lose sight of the fact that these could be withdrawn as soon as it had been militarily destroyed. RENAMO was well aware, before the cease-fire accord, that once it was achieved, the FRELIMO regime would make no more concessions. In addition, RENAMO could not rule out the possibility of an Angolan war of attrition in Mozambique.55 It was clear therefore that until RENAMO could guarantee that it had achieved enough political space to ensure its own survival within a new political dispensation, it had to maintain what Cock termed the "material base for political bargaining."56 To hold back weapons as a form of insurance was vital, in case the peace process failed or the outcome proved undesirable.

Similarly, FRELIMO had to make numerous concessions as the political price for achieving the cease-fire, nor was it sure where those concessions would lead. Unaccustomed to competitive politics, it knew that it had made many mistakes in governing, which the constituency would be judging for the first time. Moreover, the whole process meant the reallocation of the power it had been monopolising. In this sense, FRELIMO had nothing to gain and everything to lose. The question was how to guarantee the surrender of specific power, while preserving fundamental political control. To prevent this latter being jeopardised, every step should be taken, including the resort to armed means. An Angolan scenario was not ruled out by the FRELIMO side, either.

All these concerns led to the equivalent of an arms insurance policy. One of the most important moves in this direction was the rearrangement of all defence and security forces, for which no data are publicly available. What is known, however, is that from about 1992 on, there were a large number of military personnel being transferred to the National Police. The government was sure that in the event of disarmament it would be possible to protect the police, alleging maintenance of order. Furthermore, the arsenal allegedly found in the Ministry of Interior in September 1994 may have been part of this strategy.57 As a matter of fact, during the negotiation and implementation phases, police issues were non-negotiable, the government alleging that some kind of sovereignty should be maintained. There is little doubt that the FRELIMO government was preparing a force like the Ninjas in Angola.

These considerations could explain the enormous caches of weapons that have been uncovered throughout Mozambique. ONUMOZ has reported that a post-demobilisation verification mission found some 146 previously non-declared arms caches, as well as non-registered military personnel. Both were assumed to belong to the warring parties.58 Furthermore, as far as disarmament was concerned, the lack of co-operation shown by both parties when it came to the verification of arms caches outside of the assembly areas was a case in point.

Accounts from other sources support this conclusion. Smith59 pointed out that 40 per cent of the RENAMO depots have not been verified by ONUMOZ. The story about the abandoned weapons in Inhambane reported by Mosse in 1994 seems to convey intention rather than negligence.60 Lionel Dyke, director of Mine-Tech, a Zimbabwean-based company involved in mine clearance in Mozambique, declared, when interviewed by Vines: "We are finding arms caches all the time. Even in the middle of the towns ... the weapons and ammunition around here are a time bomb. Not necessarily for renewing the war, but there are plenty of guns and thousands of [rounds of] ammunition to keep criminals and poachers in business for decades."61

The Mozambican Police announced the destruction of thirty-two caches in the central province of Zambezi containing some 35 tons of weaponry in 1996 alone. What is even more blighting is the fact that the police estimate that there might be some 100 caches yet to be uncovered in the region.62 Almost every day press releases report information on arms caches in the country.63

It would be naive, therefore, to believe in the parties’ claims that those weapons caches were simply forgotten by the warring military and political authorities. A more reasonable explanation might be that most of these arms caches were deliberately hidden by the parties to the conflict during the negotiation and implementation phases of the peace process. Security and political concerns were the most important driving factors behind this stance. The weapons were kept as a guarantee: if the peace process failed or its outcome was less than desirable they could be used to resume fighting or as a political card.

Ironically, however, in the post-elections context, the arms did not have that much political value, although some RENAMO speeches may have suggested the contrary.64 In fact, it became politically ill-advised to use violence or the threat of violence to bluff one’s way into political gains. It became similarly embarrassing for the parties to admit publicly that they had held back arms during the disarmament period. Hence these weapons were left to be found by anyone, without any political connection. The parties were no longer willing to take the political responsibility for these caches. Instead they become an issue for mutual accusations.65

No matter how politicised it might be, the hard truth is that these caches contain large amounts of weaponry and constitute the biggest supply source for domestic criminals, as well as for arms traffickers and smugglers, since it is common knowledge that few, if any, arms have been imported into Mozambique in recent years.

Caches Belonging to Individual Demobilised and Active Soldiers

To consider the weaponry described above as being in the same category as weapons held by individual soldiers at their own risk and driven by personal motive would be methodologically wrong. Although in both categories security may have counted, the main reasons for retaining these arms are quite different. Individual soldiers tend to hide weapons for personal gain and to carry out acts that are criminal in nature or hold onto them for perceived personal security. This category partially meets Klare’s claim that "... once a war or insurrection has been terminated, the former insurgents often find that they have few employment opportunities in the new environment and thus turn to banditry, drug trafficking and arms smuggling".66

Reports indicate that, on being demobilised, many soldiers kept their good weapons, while declaring those of poor quality. Vines, who has spoken to a large number of demobilised soldiers, has written extensively on this issue.67 He quotes, for instance, a former FRELIMO soldier saying "... we knew that guns make good business. So we kept the best for ourselves. I have sold some to dealers from Joni [Johannesburg] and I keep others for the future. ... FRELIMO were never going to pay us for the years we were made to fight. We have to look after ourselves."68 Another informant of Vines’, reflecting on the tough life of demobilised soldiers, held that they had been betrayed by the politicians.69

It is not surprising, when all this is taken into account, that Smith estimates that there are some 40 000 demobilised soldiers who are still armed. Press releases and police reports of demobilised soldiers and other active state security personnel being found with illegal weapons are daily news. For example, in July 1995, Mozambican police seized nine guns hidden in the house of a former officer in Matola.70

Apart from the weapons that might have been kept back from the ONUMOZ disarmament exercise, there are demobilised soldiers who know where the caches referred to above are hidden. Although there may have been some deception involved, for their lobbying campaign, the Association of War Demobilised Soldiers (AMODEG) conveyed its belief that "...most of the weapons in illegal supply are in the hands of demobilised soldiers, as they know where the weapons were cached".71 In addition, RENAMO leadership, as well as some anonymous FRELIMO officers, claims that certain rashly demobilised soldiers, know the locations of the abandoned caches. Hence, whenever they are in need of a firearm, they have only to tap this source.

Moreover, there are some FADM members who may hold individual caches drawn from official armouries or from larger caches. Levels of discontent, low morale and corruption within the FADM have been reported since its inception. An initially planned 30 000-strong force, integrating half from RENAMO forces and half from FRELIMO, did not have more than 12 353 men at the time of elections in 1994. Of this number 10 per cent have since been reported absent without official leave.72

As far as recent involvement of FADM officers in the arms business is concerned, it started in the assembly areas, as the following passage, describing the safety of the stored arms, demonstrates:

"The armouries were safeguarded by a dual-key system shared between Mozambican and UN officials. When the UN left, the Mozambicans were granted sole authority over both keys. Given the poverty and levels of corruption, the opportunity was there for taking, should the camp commander want to sell some on the side. Several incidents have been reported by both Swaziland and South African police officials of Mozambican military vehicles and trucks transporting arms to locations far from FADM armouries but closer to the border."73

This account proves how supplies from official armouries have been ‘leaked’ to furnish illegal possessors of weapons. Some of the officials may also know where the abandoned caches are and extract weapons from those sources. These claims are confirmed by oral surveys conducted by the ISS Towards Collaborative Peace (TCP) project team in Mozambique. A respondent from Bela Vista reported to the team, "[W]e know there are weapons in the next town. There were a lot there some time ago but someone took them away. ... The same thing happened in the village to the south of here [apparently Pandere] and ... we do not know who has taken some of these weapons from the village but we suspect it was the authorities [police]."74 The picture becomes even clearer if this information is matched with that of another informant who was straight to the point in his accusation that "... the police sell weapons clandestinely. They have the weapons deposits."75 Police agents are also thought to be involved in leaking arms from the caches into the trafficking channels.

Furthermore, an official of the Accelerated Demining Unit (UDA), a UN programme for demining operations in Mozambique, though reluctant, also gave information that according to the agreement with the government, whenever an arms cache was found for destruction, it should be reported to the police. Surprisingly, however, within a few days of a cache’s being found and the police informed, both weapons and police seemed to disappear before any destruction could take place.76

In many instances the corruption has been officially admitted. One official police report confirmed that 102 policemen were expelled in connection with disciplinary processes. Most of those processes were related to their involvement with gangs, ranking from failure to arrest in exchange for money through the hiring of their own guns to direct participation in armed mugging.77 The most far-reaching official declaration, admitting the high levels of corruption in Mozambique and particularly in the crime-related state institutions was made by the Attorney-General in his 1996 annual address to parliament. He made it clear that in Mozambique there are crime barons or lords who are untouchable by the state.78

One could cite numerous examples of demobilised soldiers, police and active soldiers being involved in arms trafficking and other arms-related crimes. However, in summing up, the most important point to be added is this: there are weapons caches belonging to individual soldiers of the former government and RENAMO forces, as well as to the new National Armed Forces. These are weapons that were ‘leaked’ from the armouries of the two ex-armies during the war. They continue to be leaked from the armouries of the present army and police arsenals, as well as being taken from the abandoned caches. The motivation is, generally, criminal and personal gain. The weakness of logistical control during the war and thereafter; general economic hardship, especially within the demobilised community and in the FADM; and endemic corruption in the Mozambican state institutions are the main causes. Currently this category of weapons may number fewer than the previous one, but, in the long term, these may be the more difficult to eliminate. Of course, all arms in the caches and, if the leakage is not stopped, all government armouries can potentially be moved into this category.

In Mozambique, it is widely held, arms have not been used for political crimes since 1994. Hence it is from this point onwards in the possession continuum that arms start being brought into use for crime; from this point even the layman starts taking note of the arms proliferation problem. And, when arms move out of this category — if they are not in direct use here — they move into the third area, which lies at the very consumption end of the arms-possession continuum.

If arms-related crime and violence are taken as the ‘act of consumption’ of the commodity, the weapon, then this last category should be considered as the consumption aspect of the market, the final stage before the weapon is actually ‘consumed’, or comes into play, during an act of violence. In one way or the other, all illegal weapons possessed by ordinary civilians should be taken as being within the Cock ‘commoditisation of violence’ framework.79 Classification of this category of arms is difficult, for it cannot be established until a weapon has been employed in criminal action; the only way of describing a weapon’s existence is by way of its effect. Before doing that, though, it is necessary to outline briefly the main sources of the arms in this category.

Drawing from the first section of this paper, the task of identifying the sources of the weapons in the hand of civilians becomes easy. The most important of these are:

Since there were, in the first and second categories, a great number of arms eligible for supplying to criminals, traffickers and smugglers, one can expect an endless availability of weapons at this point of the possession continuum.

The sharp increase in arms-related crimes and violence in Mozambique in the last three years or so is one of the most incontrovertible facts relating to the extent of small arms proliferation. A TCP survey respondent has articulated clearly that "... there were many weapons hidden on the ground but the bandits have taken them and the result is much violence".80 Vines’81 findings show also a sharp increase in crime since mid-1993, when an unidentified gang attacked a route along a stretch of the highway linking the south and the north of the country.

Table 2
Arms-Related Crimes 1994-199682

Year	TRC	ARC	ARC as percentage of TRC
1994	26 063	1 122	4,3
1995	35 160	1 445	4,1
1996	37 725	1 679	4,4

Source: PIC—1. Policia de Investigacao Criminal

TRC — Total Reported Crimes

ARC — Arms-Related Crimes

Drawn from the general crime statistics, the data in Table 2 above are selected according to arms involvement criterion. The table includes what the police classify under the following headings: threats with arms; illegal possession of an arm; armed assault; arms trafficking; loss of an arm; arms theft.

This classification includes only those crimes which definitely involve firearms. It is clear that many arms-related crimes were omitted, therefore some qualifications should be made. First, one of the categories of crimes with a high percentage of weapons usage is car theft. To illustrate: out of forty-seven armed attacks along the National Highway No 1, in 1996, in the Inhambane province, thirty-six were car hijackings.83 Other important areas — not included in this table — in which arms-related crime frequently occurs are: homicide; drug trafficking; and physical harassment. These were not included because, within these categories, the police do not distinguish arms-related crimes from those which occur without the use of a weapon. So there was no way of distinguishing them. However, what is clear is that if figures reporting the proportion of arms-related crimes amongst these other categories of crimes were included, then the picture could change dramatically. Secondly, statistical data collection in Mozambican state institutions is, at best, disorganised, mainly because the system dealing with information relating to arms flows is very poor. Thirdly, due to the confidence deficit displayed by the police vis-à-vis the population, many crimes are not reported. People on the whole prefer not to report crimes, since to do so seems a senseless waste of time: they know that nothing will be done by the authorities. Furthermore, even in some of the reported cases, the processed information may end up disappearing before it is incorporated into reports by the department of statistics. The criminal justice system is literally paralysed.

Despite all these qualifications showing that Table 2 is an underestimation of what the picture of arms-related crimes might be in Mozambique, it does indicate the following trends. There was a considerable increase of criminal activities in general in the 1994-1996 period. While in 1994 there were 26 063 total reported crimes, in 1996 there were 37 725 criminal cases. However, as far as arms-related crimes are concerned, there were 1 122 arms-related crimes in 1994 as against 1 679 in 1996. This shows that, while general crime increased 44 per cent during the period of three year, arms-related crimes showed an increase of 49.6 per cent. As pointed out above, this picture would be increased dramatically if car hi-jacking, homicide and other crime categories, which include arms-related crimes, had been included.

Given the inherent inaccuracy of the statistics, it is necessary to complement them with other information. For example, the estimates given by a hospital care giver to the TCP survey team, that about two bullet-related casualties are daily reported to the hospital in Maputo alone,84 gives a total of 720 a year. This is nearly half of the number given by official statistics for 1996 for the entire country.

From this situation, characterised by the violent incidents of crime which followed on the war atrocities that were only too familiar, emerged a collective attitude of indifference towards firearms and death. An assault weapon in this region inspires neither fear nor anxiety: "Listen, an AK-47 is a normal thing ... here"85; "... [it] is a fact of life".86 Arms have become instruments of production, and crime and violence are normal means of livelihood and profit. This is what is meant by a culture of violence.

No information is available in Mozambique concerning the official issue of arms to individual citizens within the colonial state, before the liberation war came to an end. But, in the post-1964 period, weapons were issued to individual white settlers according to the security imperative imposed by the war. The weaker the state became, the greater the number of weapons issued to settlers to fill the security gap. Moreover, as has been discussed in the first section above, during the liberation struggle FRELIMO had, on the one hand, no motives for taking measures to control arms, because their diffusion could, in fact, help to weaken the colonial regime in general; in particular, such weapons could constitute one source of arms supply to the guerrilla fighters. On the other hand, FRELIMO, as a guerrilla movement, had to rely on popular mobilisation for the success of the war.

The Lusaka Accord which provided the framework for the cease-fire and the power transfer at independence ignored the question of weapons. It was implicitly left to the FRELIMO government to take measures in this respect in the post-independence period. Naively, or due to the regional context that emerged with Mozambican independence, the FRELIMO government failed to put disarmament on its political agenda. Here one should state quite clearly that that government had no time to take arms control measures. Shortly after it came to power, Ian Smith attacked and the war with RENAMO broke out. This train of events postponed arms control arrangements indefinitely. Instead a process of intense militarisation and arming of the whole society characterised the entire war period. A situation prevailed in which no interest in arms control mechanisms could realistically be envisaged.

As far as arms and violence are concerned, the impact of the war has been twofold: first, it not only brought many weapons into Mozambique but it also dispersed them directly to the level of the individual citizen; and second, the war created a social context in which there are many reasons for using weapons, as violence has become a means of livelihood.87

There is no doubt that all the international, regional and domestic preoccupation with arms control rests on the basis of the acknowledgement of these facts. What is being suggested here is that arms control arrangements in Mozambique are an issue that arises as a corollary to the recent peace process; an issue that occurs in reaction to the legacy of both domestic and external conflicts that led to the loss of the monopoly on the use of force by the state. Arms control in Mozambique at this stage means returning to state control the instruments of coercion that are presently spread amongst subnational actors.88 According to Rana (quoted by Cock): "Politically, the critical issue in dealing with the use of small arms in intra-state conflicts is to bring small arms back under control of the authority of the state functioning through a democratic government which enjoys broad public support."89 This is, however, a difficult task, given the fact that, "... both legal and illegal arms [possession] are embedded in intricate social relations, institutions and material interests. On the supply side, governments, manufacturers, and individual dealers will want to continue making the enormous profits involved. On the demand side ... there are strong economic interests."90

The internationally managed initiative on disarmament that was conducted by ONUMOZ as part of the political settlement of the conflict in Mozambique was the first serious attempt to control arms in Mozambique. It is important, at this stage, to discuss how the state and society at large have tackled the problem since, and to identify the major factors preventing arms control/disarmament. In analysing the mechanisms for arms control, two different levels should be distinguished: a societal level and a state level. The former includes all the activities of a civil society aimed at retrieving the means of violence from that society, in order to contribute to the enhancement of a culture of peace. The latter is an exercise undertaken by the state in its struggle to recover the means of violence diffused by war.

State Arms Control Actions

In Mozambique firearms regulations are bound to be problematic. In principle the possession of firearms is governed by law. Anyone who wants to be issued with a firearm must submit an application to the police stating why he/she needs the weapon. However, as Oosthuysen points out, "the criteria according to which a licence is granted are not publicly known."91 Given the discretional space left in the hands of the police and the lack of a tradition of accountability, the window is wide open for corruption. Once the licence for a firearm has been granted, if the applicant’s need stems from official reasons, he/she is given a gun drawn from the official armoury. If the need is personal, he/she has to import the gun, as there are neither gunshops nor manufacturers in Mozambique. Compared with the procurement of a low cost arm on the black market, this latter process seems both expensive and arduous. The result is that nobody observes the regulations, though they are in place, while the state is unable to enforce them for various reasons, one of which is the corruption mentioned.

One other problem regarding arms possession regulations in Mozambique is that, no matter whether the laws were in force before or after the war, some of the illegal arms possessors were given their arms by the government without formal issue. Oosthuysen sums up the situation in the following way: "Small arms control measures are impeded by the absence of any tradition of formal licensing of weapons in Mozambique. The government, in the past, simply distributed arms to the people. ... This suggests that many people in possession of weapons may not realise that they are now committing an offence ...."92

Bearing all this in mind, one might ask whether the government has a moral legitimacy to prosecute those who are in possession of politically distributed weapons? Or is it prepared to pay the political and/or financial price for cleaning up the mess it helped create? These are some fundamental issues the government should find answers to, if arms control regulations are to have any meaning. Otherwise, the control and collection of illicit weapons is bound to be partial only and, hence, as unsuccessful as it was in the ONUMOZ exercise.

Policing and Armouries Control

In response to a sharp increase in arms flow and violence, the government established two special task forces, one to stop highway attacks along the major routes and control cross-border gun trafficking,93 and the other to destroy arms caches. It began at the same time to seek greater co-operation with the police forces of South Africa, Malawi and Swaziland.

Since 1995, the police has been improving its strategies for combating crime, though there are no statistics to prove this. The force has undergone a process of extensive reorganisation, including the dismissal, in 1996, of Manuel Antonio and his vice-minister, popularly held as corrupt and incompetent. The numbers of arrested criminals, discovered caches and seized weapons remain daily news. For instance, Domingo,94 a weekly newspaper, reported in its 26 January 1997 issue that the police had seized twenty weapons and 181 rounds of ammunition.

The crime rates are of such political concern that the Attorney-General’s 1996 report to parliament dealt almost exclusively with arms, drugs and violence. This showed the level of awareness that crime and violence ought to be afforded in Mozambique. There is a strong force both within and outside of parliament arguing that the punishment for illegal possession of weapons should be augmented.

Although it seems that the only alternative the government envisages to arms control is improving policing and increasing penalties, which is indicated by one government official’s comments on the Christian Council’s arms buy-back project,95 alternative strategies have since been considered. According to a police officer interviewed, the special task force responsible for destroying arms caches is using an unorthodox policing approach. In fact, individuals holding arms caches, while one would expect them to be prosecuted, are co-opted, worked with and, most of the time, remunerated for the disclosure of arms caches. The rationale behind this approach is the belief that most of the cache caretakers have knowledge of more than one cache, hence to prosecute him/her would mean preventing him/her from co-operating in the disclosure of other caches.96 As a result, the Mozambican police have destroyed 9 864 weapons of different kinds seized from arms caches, as well as thousands of related accessories between 1995 and June 1998.97

A further critical move to enhance weapons collection would be for the police organisation to adopt an inward policing policy, putting its house in order by doing away with corruption before suggesting that its fellow citizens look to their houses. Indeed, no matter how effective the policing operations might be, if figures such as the unconfirmed 12 000 stolen weapons98 in 1994 alone are not properly investigated and the process stopped, then the exercise of policing becomes futile. Furthermore, analysts have noted that one of the factors for the failure of ONUMOZ disarmament was the inability of the new armed forces "... effectively [to] monitor and control the weapons which had been transferred from the assembly areas".99 Of course, lack of resources is a strong contributing factor, but an army of about 11 000 men failing to control some 200 000 weapons is, at best, a joke.

The other measure taken by the government was its quest for better co-operation with its immediate neighbours. A formal agreement on bilateral police co-operation between Mozambique and South Africa was signed in March 1995. The agreement makes provisions for the following: intelligence information on arms-smuggling exchange; access to detainees for interrogation; ground level contacts between the police forces; and joint operations. Many of the latter have taken place. Among them, four arms caches-related operations were launched, three by September 1998. Police statistics indicate that these brought about the destruction of more than 300 tons of firearms and about 40 million rounds of ammunition.

On the grounds of these successes, Rachel IV was launched. While previous operations had been undertaken in the southern part of Mozambique, Rachel IV extended to the central Sofala province. Thirty-three highly trained bomb, explosive and firearms disposal and task force operators of the South Africa Police Service and twelve of their Mozambican counterparts were employed in the operation. Despite initial setbacks, the operation has destroyed over 150 tons of illegal weapons, bringing the total figure since 1995 to over 450 tons of firearms.100

This success, coupled with the fact that arms proliferation and its impact on violence and crime have a strong regional dimension, makes regional co-operation an imperative. The technical constraints described by Potgieter,101 within the broader context of the creation of a regional peacekeeping capability, a strong confidence-building mechanism included, should guide any regional arrangement. Of course, the "armed forces, police and customs officials should be involved in collaborative training"102 and joint operations, if comprehensive arms control is to be enhanced in Mozambique and in Southern Africa in general.

Although Batchelor claims that the absence of arms embargoes was one of the factors that impacted on the disarmament process, there is no evidence for arms supplies to any of the ex-belligerent parties. On the contrary, there are many accounts that individually or in groups soldiers from both sides have smuggled weapons into South Africa,103 Zimbabwe104 (through the chimuenjes105 group) and Swaziland.106 This suggests that most of the existing weapons in Mozambique are not new acquisitions but old, surplus supplies.

For Mozambique, therefore, the most important immediate result to expect from regional arrangements is not the avoidance of incoming arms, but the removal of the link between the illegal suppliers in Mozambique and their highly demanding regional market; and hence the curtailment of the arms business in Mozambique. That is not to rule out the importance of regional efforts at controlling weapons. But what is urged here is that Mozambique should address the need to collect illicit weapons internally, as these are hindering its social and economic development. Hence, if priority must be established, disarmament should take precedence over general arms control arrangements.

Moreover, at the end of the day, disarmament depends on the warring parties’ capacity to ensure, first, that the weapons collected by a multilateral peacekeeping operation are effectively controlled in the aftermath of peacekeeping operations or in the implementing of a regional control scheme; and, secondly, that a proper policing structure, as well as other measures for further arms collection, are in place; and, most importantly, that weapons are not taken corruptly from government armouries for illicit activities.

Civil Society Arms Control Initiatives

Concern with arms collection emerged against the background outlined above and in response to the increase in armed crime in the country. Broadly speaking, the spontaneous involvement of ordinary people in weapons collection has been two-fold: either the weapons were abandoned by individuals who owned unwanted ones, but fell beyond the reaches of the disarmament exercise; or the illegal possessors of weapons were denounced by citizens to the police.

The police have acknowledged that the modest success in prevention and combating of crime is largely due to popular collaboration. While people may be predisposed to collaborate with the police, however, willingness to disclose politically hidden caches is influenced by general confidence in the twin process of peace and democratisation, as well as the prevailing economic hardship in the country. This is demonstrated by the way political and economic dynamics have influenced the Christian Council of Mozambique’s Arms Buy Back Project (TAE).

Inspired by the bible107, the Swords into Ploughshares Project, conceived in 1992 by this religious group, was aimed at retrieving the weapons known to exist in the country. The first experience of an arms buy-back programme had been attempted by ONUMOZ but aborted shortly after, apparently because of its discouraging results. Swords into Ploughshares believed that the prevailing violence in Mozambique was caused partly by the large numbers of instruments of violence in existence. Therefore, its goal was to help further the United Nations as well as government efforts to retrieve these weapons, as part of a much broader effort to enhance the culture of peace. The TAE project started officially in 1995.

Given the nature of this project, its first step was to create an atmosphere of confidence. Political actors and the police were persuaded to keep out of the way of the project so that the anonymity and security of the people handing in weapons could be guaranteed. TAE was widely accepted by government institutions, RENAMO, and national and international non-government organisations. The most vital component of the project is civic education by every possible means, the press in particular.

TAE expected that individuals would hand in their weapons in exchange for agricultural equipment and food, sewing machines and light hand machines, bicycles, construction material, as well as other material, subject to request. This meant that the initial target was to collect weapons held by individuals. The project expected to collect one weapon for each item exchanged. Collected weapons were to be destroyed and their physical constituents transformed into serviceable tools, such as hospital scissors, small tables, mugs, ploughs, cutlery and similar objects .

In February 1997, after sixteen months of activity, the project had collected some 2 016 different kinds of weaponry and thousands of related accessories, of which 881 were small arms. In August 1998 the figure was approximately 1 743 small arms. Yet, in a situation where the exact number of arms in circulation was unknown, numbers of arms collected provide little indication of the impact of the programme. TAE’s leader declares of the project’s aim: "We are not thinking about what 881 small weapons mean out of the existing arms in Mozambique, but we think in terms of the lives that would have been lost with the use of what we have collected."108

Due to the counterproductive nature of the whole disarmament undertaking, the relative success of the arms buy-back programme has been questioned.109 It is argued that the programme can be dangerous, having a negative effect on the overall objectives of disarmament by fuelling illegal arms trade. This critique assumes that the programme places itself on the demand side of the equation and hence contributes to the growth of the market. When arms buy-back programmes are taking place where other disarmament and arms control measures, such as policing, border control, arms issuing regulations and armouries control, are not effective, this is a legitimate claim. Otherwise the programme does help, no matter how little.

But in the case of Mozambique it seems that the demand for arms both within and outside of the country (especially in South Africa) did not influence the number of weapons being handed in.110 What then are the factors influencing arms collection programmes? While their overall lack of success can arguably be explained by the lack of resources, a graphic picture of such programmes, combined with those of the police collections,111 suggests a different explanation. In fact, the trend in collected weapons bears no correlation to the crime index in the country. In Mozambique arms-related crimes increased from 1993 onwards, to peak in 1997.112 In South Africa, violent crime might have stabilised but certainly it has not decreased significantly,113 which means demand has either increased or at best stabilised. Arms collected by both the police and TAE show a reality that contrasts directly with this picture. Between the start of TAE in 1995 and 1996, collected arms increased from 3 887 to 4 150. This figure decreased to 2 847 in 1997 and further to 1 186 in the first semester of 1998.114

This fact suggests that TAE-collected arms are mainly arms in the first category identified: they are more for political than personal gain and criminal purposes. This leads to a tentative consideration of the argument in which the modest success of arms collection depends to a great extent on the political momentum in the country and consequently on the peoples’ confidence in the entire twin process of peace and democratisation.

Indeed, if one relates the graph of collected arms to the political contexts throughout the period under consideration, the following correlation seems to be fairly acceptable. The general elections in October 1994 were a success. The political environment in 1995 was marked by an overall consensus on the need for joint efforts to reconstruct the country. The highest manifestation of this mood was demonstrated, among others things, by the meeting between Chissano and Dlakama115 to discuss the arms proliferation problem and the parliamentary consensus over the government’s five year programme.

Abrahamsson and Nilsson explain this positive political mood among the politicians as the "spirit of reconciliation policy".116 Whatever the reasons, this political environment seems to have determined a fairly high level of confidence on the peoples’ part in the process underway. That is especially true for those who might have held back arms for political reasons. In addition, the reintegration support scheme continued to provided something towards the livelihood of demobilised soldiers, which would certainly have guaranteed their low political and criminal profile. In such circumstances, the political value of weapons decreased and hence, ideally, whoever might have kept a weapon for political reasons was tempted to hand it in at the first opportunity. Some are even reported to have abandoned their weapons to be found by the police and the general population throughout 1995 and 1996.117 This explains the steady increase of collected weapons between 1995 and 1996.

From the end of 1996, however, popular confidence in peace and democratisation began to decline, as disruption in the political environment became evident. The final expectation that RENAMO would play a considerable political role in the country was being jeopardised by the discussion centred around the local elections. The FRELIMO government was quite relentless in forcing RENAMO to vote against the package for a local elections bill as a whole. In response, RENAMO threatened to block the elections. Subsequently, tension increased to the point where both parties’ language suggested a resort to armed action.118 The confusion surrounding the electoral process led to as high a rate of abstention as 85,4 per cent.119 At the same time, the reintegration subsidies were coming to an end. All these factors increased the political value of the arms caches once more. Hence, political control over them seemed to tighten and their caretakers became less willing to disclose them, either to TAE or to the police team. Therefore, although the police team for arms caches was reinforced by Operations Rachel, the numbers of collected weapons declined. This suggests that the politically harsh environment ensuing from the end of 1996 has been less conducive to the disclosure of politically held caches.

But an alternative explanation could be advanced here. This is the fact that the arms buy-back programme in Mozambique collects politically cached weapons. Its success depends, in fact, on the political value of the caches. However, given the fact that Mozambique has not registered serious political violence since 1994,120 prospects indicate that the political value of these caches will decline in the near future, irrespective of how tense political relations between the ex-warring parties might be. From that time onwards, economic considerations will certainly determine the decision to hand arms to TAE, the police or to smugglers. In fact, recent reports from both TAE field teams and the police team indicate that the weapons are no longer viewed as being as politically important as they were in the past. The reason for this appears lodged in the realisation that the political and strategic environment will not allow armed conflict to erupt in the country. Cache caretakers must have grasped that, although they witnessed heated political debate over the local election package, for example, the weapons they guarded were not being used as a last resort. As matter of fact, the number of caches reported both to TAE and police team actually has increased.121

This means that the political control over these caches has been loosened. Thus cache caretakers have no political reason for not getting rid of them. Recent TAE field reports indicate that, although informers claimed that they were too afraid of their political masters to disclose the majority of the weapons caches, this was merely an excuse to press TAE to provide more substantial incentives in exchange for the weapons.122 What one expects to happen in the near future is that people will tend to disclose the caches to whoever pays the highest price. While the police realised a long time ago that they needed to pay to collect the weapons,123 if arms traffickers pay more than the police or TAE, weapons from these caches will be handed over to arms dealers and, further, smuggled to wherever they are needed.

It is, therefore, of vital importance that these arms caches do not end up in the illicit market. Rather, they need to be destroyed as soon as possible. More specifically, before caretakers of caches are tempted by more lucrative illicit markets.

Recommendations

It has been made quite clear that arms flows in Mozambique are a result of two wars: the liberation struggle and the civil war. The irregular nature of these wars, and the international and regional contexts in which each took place, determined not only the excessive amounts of weapons that came into the country, but also their dispersal into the hands of the individual citizen. Above all, an attempt has been made to place the responsibility for this situation first and foremost with the two warring parties for having distributed the weapons in the first place.

It is clear, too, that no arms control measures were taken prior to the ONUMOZ exercise. Political interests led to the failure of the ONUMOZ disarmament process. As a result, Mozambique is literally awash with small arms. The impact of this is to be seen in the increase in violent crime, mostly using firearms. It is widely recognised that if arms in the possession of criminals are not collected with some urgency, not only Mozambican, but regional stability and development will be in danger.

Most of the information available suggests that these weapons are not being procured elsewhere, as was the case in the past, but are remnants of the war, circulating either from caches or the official armouries to the individual criminal, trafficker or smuggler, operating throughout the region. Furthermore, economic deprivation and general destruction of the social fabric certainly provide ideal conditions for the arms business.

In the light of these facts, the following recommendations should be implemented with urgency:

* The author wishes to acknowledge the valuable assistance given him by Virginia Gamba and Sarah Meek throughout his writing of this monograph. Further, he wishes to thank all the members of the Institute for Security Studies for their hospitality, which created an environment such that only he could be held responsible for any deficiency in his academic performance.