Chapter 3
ANGOLA'S INSTITUTIONAL CAPABILITIES FOR DD&R:
EXPERIENCES AND LESSONS LEARNED FROM PREVIOUS PROCESSES

---

Published in Monograph No 83, April 2003

Sustaining the Peace in Angola
An Overview of Current Demobilisation, Disarmament and Reintegration

João Gomes and Imogen Parsons

While this is the first time since their displacement that large-scale return of IDPs has taken place in Angola, the current disarmament, demobilisation and reintegration process is the third in Angola’s history of civil war, following the failure of the 1991 Bicesse Peace Accords and the (temporary) collapse of the 1994 Lusaka Protocol. The manner in which each was implemented (or not) had radical implications for the unravelling of the peace process each time, as well as for the local stability and security of regions around Quartering Areas. Furthermore to a large extent, the current DD&R process builds on the institutions and may benefit from lessons learned from these two previous attempts. As a consequence a brief discussion of these seems appropriate.

The Bicesse DD&R process

The first comprehensive peace agreements between the government of Angola and UNITA, the Bicesse Peace Accords signed in May 1991, contained the origins of the DD&R structures still in place today. Moreover, the accords stipulated the quartering of both UNITA and government troops within 60 days, the creation of a 50,000 strong joint armed forces and the demobilisation of surplus UNITA troops. From an estimated total of 200,000 men in arms on both sides, the new Angolan Armed Forces (FAA) would be composed of 20,000 soldiers from the FAPLA, the government’s army, and 20,000 from FALA, UNITA’s army as well as by an Air Force of 6,000 and a Navy of 4,000. On a political level, the realisation of multi-party elections in September 2002 was made dependent on the formation of the Angolan Armed Forces while completion of disarmament and demobilisation were not insisted upon, however. In retrospect, this was a mistake which aided the return to war, even if it cannot be said to have caused it.

In institutional terms, the Bicesse Accords established three ‘joint commissions’ in order to oversee and verify the process, built around the Joint Political-Military Commission (CCPM)²⁸, which was composed of the Government and UNITA, with the Troika (Portugal, Russia and the United States) solely as observers²⁹. A new inter-sectoral body, the Interministerial Office to Support the Demobilized Military of Angola (GIAMDA), was created on 15th of November 1991 by presidential decree with the aim of establishing a broad programme for reintegration of former combatants. Supporting structures in the provinces were then established to co-ordinate the DD&R process. The role of external actors other than the Troika was extremely limited. The signatories of the Bicesse Accords decided to extend the United Nations’ presence in Angola, UNAVEM I³⁰, under a renewed mandate. In response to a formal request from Angola for United Nations’ participation, the Secretary General reported to the Security Council and proposed that the existing mandate of UNAVEM I be expanded to include verification of the cease-fire and neutrality of the Angolan police. On 30 May 1991, the Security Council adopted resolution 696 establishing UNAVEM II.³¹ While having a new mandate that included overseeing implementation of the various provisions of the Accords, including organising the elections, responsibility for compliance lay firmly with the two parties. As a result, UNAVEM II’s power to intervene in the process was severely limited, being restricted to a supervisory role.

In addition, the United Nations’ abilities were limited by the lack of adequate resources, with an initial budget of only $132m for a seventeen-month period and a staff component of only 350 military observers and 126 police observers. The extent to which UNAVEM II had limited resources for the tasks at hand can be assessed in comparison with the UN presence in Namibia, UNTAG, which cost $383m for a 12 month period, in a country with a significantly smaller population. In fact, the ratio of UN personnel to population for UNTAG was 1:150, whereas for UNAVEM II it was 1:16,000.³² Margaret Anstee, United Nations’ Special Representative during the Bicesse process, has famously compared this lack of resources to being asked to fly ‘a 747 with the fuel for a DC-10’.³³ In another occasion, Anstee added that ‘adopting a myopic approach to UN peacekeeping, the Security Council gave UNAVEM II a mandate and resources that absurdly underestimated the enormity and complexity of the tasks that lay ahead…UNAVEM II was a misguided exercise in peace-keeping minimalism’.³⁴

As a consequence, the disarmament, demobilisation and reintegration process under the Bicesse Accords did not proceed smoothly. Although rapid demobilisation was called for, the accords did not allow sufficient time for proper planning and implementation. In fact, partly a consequence of technical failures in the accords, particularly as regards independent monitoring activities, partly a result of the intransigence of both parties, secret armies were maintained in violation of the Accords.³⁵ Quartering and registration of both government and UNITA troops was never completed and it is not known exactly what UNITA’s fighting strength was at the end of the war, since, although it claimed force levels in the region of 75,000, this number may have been inflated for strategic reasons. In addition, of those that did register, a significant number were believed to be outside normal fighting age as well as unarmed.³⁶ On the government side, a significant number of troops either ‘self-demobilised’ after registration or went AWOL. In fact, over 12,000 government troops are believed to have gone missing from October 1991 to February 1992.³⁷

As regards disarmament, an important achievement during the Bicesse negotiations had been the acceptance by both sides as well as the Troika of observer countries of the so-called ‘triple-zero’ clause in which the parties agreed on restraining from acquiring lethal material, and the United States and the Soviet Union agreed to cease all supplies to any of the parties and encourage other countries to do likewise. However, as relates to existing weaponry, significant numbers of UNITA troops reported to the quartering areas unarmed. The slow rate of disarmament during the Bicesse process may be an indication of both a lack of seriousness or confidence between the belligerents as well as a reluctance to hand over weapons which could be sold for profit in a situation characterised by a lack of employment opportunities.³⁸ Due to the lack of resources for infrastructure development, a ‘double-key system’ was never introduced in Angola (which could have guaranteed that arms were stored in a safe place, one key remaining in local hands and the other with the United Nations), contributing to the difficulties in verifying disarmament. In fact, on the government side many of the troops that ‘self-demobilised’ took their weapons with them when they left, resulting in a rise in incidents of banditry and crime in particular in areas around the camps.

Demobilization of troops did not, in the end, begin until almost a year after the signature of the Accords, in March 1992. Furthermore, although the parties revised the initial projection of their combined total troop strength to 151,930, by 12 February 1992 only 61% of troops were quartered, corresponding to 50.44% of government troops and 93.75% of UNITA troops. Because quartering and registration had not been completed, of the 25,000 UNITA troops scheduled for demobilisation only 10,402 had been demobilised by September 1992 when elections were due. On the government’s side, 123,887 FAPLA troops had been demobilized in total. Those soldiers that were demobilised received a package including money to the value of about $100 (equivalent to five months salary) and a clothing kit. UNITA combatants were demobilised directly, without first being incorporated into the existing FAPLA since the objective was the creation of unified Angolan Armed Forces.

As regards reintegration programmes, although GIAMDA had been created with the intention of creating a multi-sectoral social reinsertion programme for demobilized soldiers, including financial support and professional and vocational training, all plans and programmes subsequently crumbled when, in October 1992, Angola returned to war. In fact, reintegration programmes and schemes approved by the CCPM, and amounting to $447 million, were never implemented. As Margaret Anstee points out, ‘slow demobilisation and the formation of the new armed forces led to the spectre of there being not one, but three armies in place at the time of the election. This combined with the proliferation of weapons among the civilian population, the decline in law and order, and the slow progress in extending the central administration, added up to a very grim prospect’^.39 Nevertheless, because the formation of joint armed forces was one of the pre-conditions for elections, the parties went ahead and nominally created the FAA on 27 September 1992, just two days before elections were held. The experience of Bicesse was a painful one, but with many important lessons, as can be seen in the box below. Some of these were born in mind in the design of the 1994 Lusaka protocol, although with little more success.

The Lusaka Process

When the Lusaka Protocol was signed in November 1994, after two years of high intensity conflict in Angola, a new framework for disarmament, demobilisation and reintegration was developed. Moreover, the drafting process had taken note of some of the failings of Bicesse as discussed above, and attempts were made at remedying them. In fact, the Lusaka Protocol reaffirmed the Bicesse Accords setting forth the details of a cease-fire, a second round of presidential elections, demilitarisation, disarmament and the formation of a unified armed force and national police force. The United Nations was involved in the negotiations, and an expanded mission, UNAVEM III, was created by Security Council resolution 976/1995 in February 1995. The timeframe for the process was more elastic, with no firm date set for the 2nd round of Presidential elections to take place. In the meantime, a power-sharing system was devised, whereby members of the government and UNITA would form a Government of Unity and National Reconciliation (GURN), where Savimbi would be accorded special status as leader of the largest opposition party. It was 1997, however, before the GURN came into being.

Box 1: Lessons learned from the Bicesse DD&R process Sufficient time for proper planning and implementation of the process is essential—the timetable for the Bicesse process was too rigid. In the headlong rush to demobilise the very important quartering and disarmament processes were neglected and UNITA were able to return easily to war. While this is not now considered a threat, the importance of devising a sufficient (and realistic) timeframe is clear. Elections should not be rushed, especially based on a ‘winner-takes-all’ model. In 1992 elections merely continued the polarised military logic of the war, and gave no way to deflect potential ‘spoilers’ to the process. This political/military polarisation and the potential for revival of conflict must be born in mind when elections are scheduled. Attention must be paid to conditions within quartering areas and among demobilised soldiers. Under Bicesse, lack of pay and poor conditions led to a mass self-demobilisation of government troops, causing heightened insecurity around the camps. While UNITA troops have proved more disciplined, this is nevertheless still a risk. Sufficient resources must be allocated to support a DD&R process. UNAVEM II was severely limited in its mandate and resources and unable to enforce compliance with disarmament and demobilisation, and furthermore the United Nations felt itself disadvantaged since it had not been involved in the process of drafting the accords. This also indicates that participation in the process of negotiation and drafting accords is important in ensuring their success. Programmes and projects for reintegration must also be taken seriously and are an integral part of DD&R. Under Bicesse insufficient attention was paid to questions of reintegration and UNITA were not sufficiently involved in the consideration and design of programmes, subsequently resisting the intervention of GIAMDA.

A number of new institutions were created to oversee and implement the DD&R process. In place of GIAMDA, in March 1995 a new body was created, the Institute for the Socio-Professional Reintegration of Ex-Military Personnel (IRSEM⁴⁰), under the Ministry for Reinsertion and Social Assistance (MINARS). A Joint Commission was again formed, consisting of Government and UNITA as well as a representative of the United Nations, Alioune Blondin Beye who had been the principal mediator of the Lusaka Protocol. A technical working group was created to supervise and advise on the disarmament, demobilisation and reintegration process, as well as in the creation of the joint armed forces, or Angolan Armed Forces.⁴¹ Meetings of this group were also attended by UN agencies and NGOs as appropriate to the discussions, allowing a broader dialogue to take place.

UNAVEM III had a two-year mandate extended beyond that of UNAVEM II to include a formal role in overseeing the demobilisation and integration of the armed forces, and responsibility for decreeing when the conditions were appropriate for elections to be held. It was given over 7,000 peace-keeping troops in addition to 350 military observers and 260 police observers. Consequently, UNAVEM III had an expanded role within the DD&R process, with representatives and peace-keeping troops in each assembly area as well as responsibility for troop registration and disarmament, camp co-ordination and production of demobilisation documentation. United Nations’ agencies as well as various NGOs were also present in many assembly areas and assisted in the process.

Although expected to complete its mission and leave Angola by February 1997, UNAVEM III was later given a phased withdrawal, as it became clear that the process was not proceeding as smoothly as had been anticipated. In fact, at a cost of as much as $1million a day, UNAVEM III was deemed unsustainable and ineffective and peace-keeping troops were gradually withdrawn. As of July 1997, UNAVEM III was replaced by the more modest United Nations Mission of Observers in Angola (MONUA). When war officially resumed at the end of 1998, the United Nations Security Council announced the termination of MONUA’s mandate effective from February 1999. In October that year a new presence was established, the United Nations Office in Angola (UNOA), with a mandate ‘to explore effective measures for restoring peace, assist the Angolan people in the area of capacity building, humanitarian assistance, the promotion of human rights and coordinate other activities’.

Despite the knowledge that inefficient quartering and demobilisation under Bicesse had been a factor in the resumption of war, the process was scarcely better handled this time. According to Alex Vines, ‘in 1992 one of the main problems of the peace process was UNITA’s failure to demobilise most of its fighters. During the Lusaka process, the quartering and reintegration process was also slow. The operation only started in earnest in February 1996. It was incomplete and involved few key UNITA troops; conversely many in camps were civilians’.⁴² The timeframe did indeed allow increased flexibility, but it has been argued by some that this was taken advantage of by UNITA in particular, allowing them to regroup and rearm, and in fact contributed to the resumption of war in 1998.

As regards quartering and registration, the Lusaka Protocol established a minimum number of soldiers to be quartered by UNITA, some 62,500. They were to be gathered in 15 Assembly Areas, later renamed Selection and Demobilisation Centres, where UNAVEM III would assume responsibility for disarming them. There were a further 5 War-Disabled Centres, which were eventually included in the demobilisation process for logistical reasons. Each assembly area contained a United Nations Administrative Centre and United Nations Troops, and a representative from UCAH who was responsible for the co-ordination of camp management, registration of UNITA soldiers and production of demobilization documentation. Included in the registration process was the collection of socio-economic data such as family size and intended destination. Fortnightly head-counts were used for monitoring of soldiers’ movements and for updating information.

Nevertheless, the process proceeded slowly, with UNITA suspending the process in December 1995 in protest at the government’s seizure of UNITA-held territory. In early 1996 diplomatic pressure was applied by the Portuguese President, Mario Soares, and US Ambassador to the United Nations, Madeleine Albright, with the result that Savimbi promised to increase the pace of quartering and there was a temporary increase in the number of troops registering.⁴³ By July 1996, when the new joint army was officially sworn in, 70,600 had been quartered, but it is believed that many were not in fact soldiers but had been conscripted to make up numbers, and a high proportion were underage or disabled (4,799 and 10,728 respectively, barely short of the eventual totals demobilised). Registration and disarmament of residual UNITA personnel finished on 22 December 1997, with a total of 78,886 registered, of which 8,607 were underage and 11,051 disabled. Of these around 26,000 ‘deserted’, either during the headcount or demobilisation itself.

Disarmament under the Lusaka Protocol was a critical issue, perhaps more so than during the Bicesse process. In fact, when war resumed in 1992, both parties abandoned the ‘triple zero clause’ with the result that weapons procurement sky rocketed on both sides.⁴⁴ Even so, disarmament of UNITA troops under Lusaka was a highly ineffective procedure, with a high proportion of weapons unserviceable or in poor condition. No heavy weaponry was handed in, and it was clear that UNITA was maintaining significant arms supplies. As part of the Lusaka Protocol, the government was also required to disarm the civilian population it had distributed weapons to in 1992. As fighting resumed following UNITA’s rejection of the election results, anything up to a million AK-47’s were handed out in Luanda alone, and few had been recovered. In August 1997 the government suspended this process, however, citing UNITA’s reluctance to disarm. The process was to resume once the normalisation of state administration had been completed, but this never happened. Furthermore, despite the stipulation in the Lusaka Accords that neither side was to rearm (Annex 3, Section IV) the government received weapons including tanks and artillery in May 1995, believed to represent purchases made since Lusaka.⁴⁵ UNITA similarly continued to receive supplies and weapons via Zaire and Congo-Brazzaville in particular.

Although demobilisation had been planned in two phases, first the demobilisation of underage soldiers and then of adult soldiers (both able-bodied and disabled), in practice the two frequently took place side by side. The majority were demobilised under a Rapid Demobilisation Plan finally agreed upon by the Joint Commission after a number of drafts and revisions in 1996 and early 1997. While the Lusaka Protocol had specified a number of changes from Bicesse in order to minimise the risks associated with rapid demobilisation, a number of these were now deemed impractical and abandoned. It had initially been decided that all UNITA troops to be demobilised would first be incorporated into the Angolan Armed Forces, and the precise numbers for incorporation or demobilisation determined later. In the event, surplus UNITA troops were demobilised directly, without passing through the FAA. And, although the normalisation of state administration was also to have been completed before demobilisation could begin, when it became clear that this was not practical, a phased plan was introduced beginning in the areas around the quartering areas.

The ‘second phase’ of demobilisation, covering adult troops but in practice also many underage soldiers began in April 1997, in Vila Nova, and continued into 1998. The first phase had covered almost 2,484 troops, including 360 underage soldiers from the Popular Armed Forces for the Liberation of Angola (FAPLA). By May 1998 a total of almost 50,000 UNITA troops had been demobilised, of which 5,059 were underage and 10,771 disabled. This was out of a total of 76,360 troops registered as eligible for demobilisation, however, after the incorporation of 10,880 into the FAA, meaning an excess of over 25,000 remained. The joint army was sworn in on 10 July 1996.

The process was beset with difficulties, with demobilisation taking place on the basis of a final headcount, two weeks before ‘demobilisation day’. Soldiers not present were to be considered deserters and not eligible for benefits, although this rule was relaxed in some places. It seems that after the repeated delays many had not really believed this time it was real, and were instead working in the fields they had begun to cultivate around the camps.⁴⁶ Around 26,000 UNITA soldiers were eventually registered as deserters. UNITA frequently interrupted and delayed the process, sometimes simply refusing to go ahead with demobilisation. At the same time reports emerged that UNITA was regrouping, forcibly recruiting and even training in Jamba. Attacks frequently put down to bandits escalated in 1998, and human rights abuses and revenge attacks by both sides were common.⁴⁷ In addition, the resettlement and reintegration of demobilised soldiers was a complex procedure, involving an array of government, UNITA, UN agencies and NGOs. This helped to address concerns about the government’s capacity to implement such a complex programme, but the difficulties of co-ordination frequently led to delays and confusion. Upon leaving the camps soldiers were issued with travel cards (by the International Organisation for Migration, IOM), demobilisation cards (FAA), benefit cards (Unit for the Coordination of Humanitarian Affairs, UCAH), subsidies (IRSEM), reintegration kits (IOM), a ‘Portuguese kit’ supplied by the Government of Portugal, a World Food Programme food ration and assistance in returning to a destination of their choice. This assistance was provided to 40,631 UNITA ex-soldiers and 107,197 dependents, and 4,181 resettlement kits were also provided to families of UNITA troops incorporated into FAA.

There were problems in implementation however. Since resettlement assistance was classified as a benefit, and not as a right, soldiers were not able to change their minds about their destinations after the final headcount, unless the destination happened to fall along the same route. On the day of demobilisation many soldiers attempted to change their destinations, often seemingly under coercion from UNITA.⁴⁸ Payment of ‘special subsidies for assistance and reintegration’, SEAR) was a further problem, beset with difficulties such as bad management of funds coupled with fraud. At the end of 1998, when war officially resumed, only around 60% of demobilised soldiers had received the second payment and only 25% the third and final payment.⁴⁹

Reintegration programmes were designed on the basis of socio-economic surveys carried out by UCAH following Bicesse and also in the early stages of Lusaka. These gave a profile of the ‘average soldier’ as being ‘a man between 25 and 30 years old, with 4–5 years schooling, literate and able to speak Portuguese. Although from a rural background, where he or she attended school, many had travelled to the city prior to military service. This mobility continued in the army. During the 8–12 years of service in the army, the average soldier would have been stationed in various provinces and have come into contact with people of various ethnic backgrounds’.⁵⁰ This survey was related only to Government troops, however, and the first survey to cover both Government and UNITA was published in 1995.⁵¹ Respondents were asked about their length of service, desire to demobilise, intended future destination and occupation. When asked to prioritise reintegration programmes, the first priority of both FAA and UNITA was training and education (43% and 52% respectively). UCAH warned, however, that ‘the myth that training and education can solve all problems that demobilised soldiers will face during reintegration is slowly becoming established and could possibly lead to the creation of unrealistic expectations not only among demobilised soldiers but also among Government Officials, UN agencies, NGOs or any other organisation involved in the demobilisation process’.⁵² The socio-economic profile of ex-combatants during the Lusaka process can be seen below:

Box 2: Ex-Combatant socio-economic profile during the Lusaka Process (sample of 160.000) Average age: 26 8% below 18 years 56% below 25 years 27% did not have any education Only 12% had more than primary school (4th form); 2% had more than 6th form Average families: 6 people 27% did not have families 60% had three family members 89% from rural areas (Huambo 23%;, Bie 19%, Benguela 13%, Malange 12% and Huila 10%) 85% demonstrated willingness to return to rural areas Average 8 years in the army (1/3 of the life of the soldier) 55% were soldiers for more than 5 years 40% want to become farmers; 21% students, mechanics or carpenters; 11% motorists; 7% health technicians.

However, training did form the backbone of the plans for reintegration. Three major training actions were agreed upon. These were: the ‘Community Counselling and Referral Service for Demobilised Soldiers (SeCoR)’; the ‘Training for Self-Employment for Ex-Military Personnel from Angola (TSE)’; and finally, the ‘Community Based Quick Impact Project (QIPs)’. SeCoR and QIPs were the responsibility of UNDP while TSE was to be handled by UNDP and ILO. The majority of projects never materialised, however, due to the resumption of war in 1998. An additional alternative strategy was the creation of ‘work brigades’, known as the ‘Quarto Ramo’, or fourth brigade of the Armed Forces. This was dedicated to reconstruction projects such as roads and bridges, and intended to provide economic and practical training opportunities for demobilised soldiers as well as providing a readily available workforce for construction and infrastructure projects of this kind. It was an expensive project, however, and the political implications for UNITA of having its people seemingly degraded in this way were too great for Savimbi to agree to it.

In conclusion, although some of the lessons learned during the Bicesse process were incorporated into the design of the Lusaka Protocol and the subsequent demobilization and reintegration processes, a number of mistakes were made. The lessons learned during the Lusaka process can be seen in the box below.

Box 3: Lessons learned from the Lusaka DD&R process While some flexibility and time for planning is necessary, once agreed timetables should be followed as closely as possible. The schedule for completion of quartering, disarmament and demobilization under Lusaka was too loose, with the effect that UNITA were able to delay the process and regroup and rearm in the meantime. While this is no longer a serious threat, planning and implementation of DD&R programmes are dependent on agreed timetables being followed. Delays also have the effect of weakening confidence on both sides. Under Lusaka delays and irregularities in the payments of benefits undermined confidence in the process among demobilizing UNITA soldiers in particular and lessened their inclination to co-operate. The loss of confidence and co-operation on the part of demobilising soldiers may create serious security risks, especially in a heavily armed society. Again, sufficient resources are essential to accompany any DD&R process. Although better equipped than under Bicesse, UNAVEM III still did not have the means to effectively monitor disarmament or to prevent UNITA from rearming and regrouping. While this is not considered a threat post-Luena, successful implementation of DD&R requires the commitment and rapid mobilisation of funds. The extension and normalisation of state administration is essential to consolidate peace and ensure security. The failure to do this under Lusaka worsened the situation, heightening insecurity, complicating demobilization and obstructing reintegration programmes. In the context of mass return of IDPs and ex-combatants this must be a priority. Confidence on the part of leaders of both sides, but also the ex-combatants themselves, is a pre-requisite for successful DD&R. The entire process was conducted in an atmosphere of mutual suspicion and distrust, although it is difficult to see how trust could have existed between the two, given the events of 1992. Nevertheless, confidence in the outcomes of the process must be built on a mutual basis and must extend right down to the individual ex-combatants themselves. Demobilisation and re-integration are a ‘family affair’: the provision of basic services to soldiers’ dependents should be considered an integral part of the peace process and provided for in the peace agreement. The funding for these services should come from the assessed budget, and any support provided to assembled troops should also be provided to their dependents.53 Re-integration is a ‘community affair’: community-oriented re- integration programmes need to be evaluated to determine their effectiveness and provide guidance for future re-integration programming.54