Rebels Create Havoc in Mozambique


In December 1987, Robert Gersony was engaged by the Department of State's Bureau for Refugee Programs to undertake an assessment of designated Mozambican refugee matters. The bureau's director, Ambassador Jonathan Moore, decided to have the assessment conducted as a result of the bureau's perception of a mounting refugee crisis in southern Africa.

The bureau had witnessed an increase of 300 percent in the number of Mozambican refugees in southern Africa over the past year. The bureau currently estimates that the total number of such refugees is about 870,000. Malawi, for example, has 450,000-500,000 refugees - up from 70,000 just 15 months ago. Reports from the field continue to indicate that 20,000-30,000 Mozambican refugees per month - up to 1,000 each day - enter Malawi. Many arrive in poor health, severely malnourished, without belongings and often naked. In Swaziland, there were about 8,500 Mozambican refugees at the beginning of 1988; the number has already doubled and, according to one informed estimate, could quadruple by midyear. Increased flows into Tanzania, Zambia, Zimbabwe and South Africa combine with Malawi and Swaziland arrivals to suggest that there will be more than one million such refugees by year's end.

Complementing this exodus is a high rate of population displacement within Mozambique. According to the Bureau for Refugee Programs' World Refugee Report (1987), more than one million Mozambicans are internally displaced.

Gersony was engaged by the Bureau for Refugee Programs to shed additional light on such issues as the causes of these refugee flows, the likelihood of continued migrations, refugee protection and assistance and the possibility of refugee repatriation.

To do this, he conducted a field visit of nearly three months to 42 different locations in five countries - Mozambique, Malawi, Zimbabwe, South Africa and Tanzania. In 25 refugee camps, separated by as many as 1,500 miles, he spoke at length and individually with nearly 200 randomly selected refugees. Forty-three percent of them had arrived at these camps during the period of the assessment, some within hours of the interviews conducted for this study. They came from 48 different districts of Mozambique, including northern, central and southern provinces. Some came from areas of predominant Frelimo (government) control, others from areas of Rename (Mozambique National Resistance Movement) prevalence. At many of the 25 refugee sites and in 17 other locations visited, independent national and international religious and relief assistance workers were also interviewed and added valuable complementary information.

This report attempts to recapitulate the complex body of information gathered from refugee and other accounts. In order to convey this information, the collective accounts have been divided into nominal categories. The report is divided into three sections: Assessment Procedures, Conflict Dynamics and Summary of Findings.


Geographical Scope

The Mozambique refugee assessment was conducted during a period of approximately 11 weeks between January and March 1988 in 52 locations in five southern African countries: Mozambique, Malawi, Zimbabwe, South Africa and Tanzania. Roughly 25 sanctuary sites for refugees and displaced persons (10 sites in six of 10 provinces in Mozambique; 15 in the neighboring countries) were visited. In addition, the five national capitals and 12 administrative centers in which hospitals and relief operations are based were visited.

Source of Findings

The principal source of the assessment's findings was 196 individual interviews with refugees and displaced persons who provided eyewitness accounts of the incidents and patterns of conduct in the conflict which led to their migration. In addition, the author met with roughly 50 priests, ministers and religious leaders and national and international relief workers with independent, firsthand experience in assisting the refugees. They provided background information and an opportunity to check aspects of the information provided by the refugees. In the national and regional capitals, the author met with host government, US Embassy and other diplomatic officials.

Refugee Interviews

In each of the 25 locations in five countries, the author as randomly as possible personally selected or directed the selection for interviews of the refugees and displaced persons (hereinafter "refugees"), without reference to the type of conflict problems they might have experienced in Mozambique. He also attempted to select refugees who had arrived as recently as possible and who came from as wide a geographical distribution as possible.

The interviews were conducted in the language which the refugee identified as his or her native language, through translators selected by the author. In the three countries that accounted for about 75 percent of the interviews, the author was able to select, from among the agricultural development staff of the US Agency for International Development (USAID) missions, an individual who was a native speaker of one of the principal languages spoken by the refugees. This individual assisted in translation of the interviews in all of the sites in his country. In most other cases, local workers selected by the author who are associated with health clinics, religious and relief organizations or local relief committees assisted in translation on an ad hoc basis.

The average length of each interview was one hour; the more detailed ones took as long as two hours. The interviews were conducted in a private place, out of hearing and usually out of sight of those not participating in the interview. In Mozambique, those present in each interview were the refugee, the translator(s) and the author. In some of the other countries, a local official was also present. Sometimes the refugee's spouse and children, especially for those who had just arrived at the sanctuary sites within minutes or a few hours of the interview, were also present. For a list of interview sites, see box.

Refugee Arrival Dates

In order to obtain the most current data, preference in selection of refugees to be interviewed was given to those who had arrived recently. Nearly 90 percent of those interviewed had arrived at the refugee site during 1987 and 1988; about 50 percent of these were January-March 1988 arrivals. The 10 percent who arrived before 1987 also provided useful information, but more than one-third of these had migrated earlier because of drought or in order to seek employment in the more prosperous neighboring country; thus, the data for 1987 and 1988 arrivals is the most useful.

Districts of Origin

In order to obtain a wide geographical representation among the refugees, preference in selection was given to those from different home districts and, as appropriate, from different villages within such districts. The 196 refugees interviewed reported that they came from villages in 48 different districts of Mozambique. These districts were widely distributed in Mozambique, with each province in the country represented to varying degrees.


The majority (65%) of the refugees interviewed described themselves as farmers. Almost all were residents of small villages and hamlets. The occupational breakdown of the respondents is as follows:

Farmers 65%

Farm workers 4%

Tradespeople/Merchants/Day laborers (urban) 9%

Teachers 5%

Students 5%

Health workers 2%

Housewives 2%

Children 2%

Former soldiers 2%

House servants 1%

Miners 1%

Other 2%

Previous Mobility Affecting Migration

Of the 55 Mozambicans interviewed in displaced persons sites inside Mozambique, 40 percent had never traveled outside the country. About 45 percent had visited one of the potential asylum countries neighboring Mozambique earlier in their lives.

Of the refugees interviewed outside Mozambique, 45 percent reported that they had at least once before visited the country in which they are currently in asylum. But 50 percent reported that when they migrated from Mozambique to the interview site, it was their first trip outside the borders of Mozambique.

Refugee Accounts

The refugees often provided detailed eyewitness and other credible accounts, spontaneous responses to follow-up questions and simple diagrams drawn in the dirt to illustrate. The majority of the respondents had no formal education or literary skills, but they described effectively what they had seen and what had happened to them. A few refugees appeared afraid to provide any information. The great majority, however, appeared to speak forthrightly and with confidence about their experiences.

The refugee reports tended to be consistent throughout the 25 interview sites in five countries, with origins of the refugees, as well as their sanctuary sites, separated, in some cases, by as much as 1,500 miles. In addition, a number of the reports tended to corroborate each other, with refugees inside Mozambique describing the same events in the same places as had been described by refugees in camps outside Mozambique. Certain additional specific opportunities for cross-corroboration arose. In one asylum country, several refugees described a series of incidents within Mozambique in considerable detail. When the author visited Mozambique, he was able to visit the area discussed and to verify the information they had provided. The major difference between their description of what had occurred and that provided by survivors who had remained in Mozambique, and separately by local officials, was that the refugees had given a lower estimate of the number of persons who had been killed or wounded. Refugees provided sufficient detail so that accounts by several refugees - sometimes in different countries - could be readily identified and thus not "double-counted" in the quantitative analysis.

Distinguishing Between the Parties

Serious complaints about the conduct of both sides of the conflict were received, although, as the report will describe, the number of complaints about one party exceeded by a considerable proportion the complaints against the other. Complaints against each side tended to contain the same level of detail and authority and to be conveyed with the same level of conviction. There were no complaints against the government (Frelimo) soldiers from refugees inside Mozambique. All of the complaints against these came from refugees outside Mozambique. This may reflect a natural hesitation of refugees in the government-controlled areas to express criticism of the soldiers under whose protection they live. This constraint may have affected the 28 percent of the sample interviewed in Mozambique. But criticism against the government soldiers received from the refugees in the four neighboring countries representing 72 percent of the sample was also very low. These included two countries considered either hostile to the Frelimo regime or, until recently, considered more neutral in the conflict. In short, this factor is unlikely to have significantly affected the findings of the assessment. Soundings among the independent religious and relief assistance sources tended to reinforce this conclusion.

The refugees spoke authoritatively about their ability to distinguish between combatants of the two sides in the conflict. Many of them were questioned closely about the basis for their assertions and explained their views effectively. They pointed to personal familiarity with the individuals in the government forces, who were often stationed in or near their villages; to the heavy casualties often suffered by the government side in the attacks by insurgent forces on villages. In some areas, some refugees described marked differences between the two sides in physical appearance, including hair length and style, and face and body adornment. Routine patterns of conduct and other data provided by refugees tend to confirm their assertions. Refugees were also asked whether freelance bandits, as opposed to elements of the two principal parties to the conflict, perpetrated the acts they witnessed; their responses were emphatically negative. When in doubt about the identity of assailants, they seemed prepared to say so.

Moreover, one-fifth of the refugees interviewed said that they had involuntarily resided for many months - sometimes for years - in areas where insurgent influence prevails. They had often been used by Renamo (the Mozambique National Resistance Movement) as porters to carry materials over long distances within these areas. Thus, a significant proportion of the refugees possessed not only the ability to distinguish between the two sides, but also a firsthand knowledge of the operating procedures of both forces.

The majority of the refugees lacked two skills that limited the precision of their data. Most could not count accurately above the number 10; as a result the number of deaths appeared to be underestimated. Beyond the number 10, these refugees could only describe "many," and could not assign a corresponding number to the concept except to say that it exceeded 10. The second limitation was an inability to identify the month and year of an incident. Such identification had to be established approximately, by working back through seasonal agricultural thresholds, such as the advent of rain, planting and harvest. Neither limitation materially affects the conclusions of the assessment, however. In quantifying casualty reports the term "many" was almost always assigned a maximum value of 10.

In summary, the author would assign a high level of credibility to the refugee accounts.


The 1987-1988 Arrivals

Nearly 90 percent of the refugees interviewed had arrived at the refugee site where the interview was conducted during 1987 and 1988, about half within the three-month period of January-March 1988. More than 90 percent of these 1987-1988 arrivals said that they fled their home areas because of the conflict, specifically because of abusive conduct by the parties to the conflict. About 5 percent of this sample said they had migrated because of drought, 2 percent to seek employment and 2 percent for other reasons. The pre-1987 arrivals represented only about 10 percent of the sample; one-third of these said that they had migrated during earlier periods because of drought, in search of employment or for other reasons not directly related to conflict violence. In order to provide conclusions that are up-to-date, and to avoid undue emphasis on the drought and economic migrants of earlier years, data presented in the following sections will be for the 1987-1988 arrivals unless otherwise specified.


Parties to the Conflict

Each of the refugees who fled from their homes because of conflict (more than 90 percent) attributed their flight to one party or the other or to both parties in the conflict. The clearest way to summarize the experiences of these nearly 200 refugees as well as other information collected in the assessment is to describe the conduct of the war as the refugees themselves have experienced it, separately for each side: for Renamo, the insurgent forces; and for Frelimo, the defending government forces. The proportion of attention devoted in the report to each of the two respective groups reflects the proportion contained in the accounts provided by the refugees themselves.

Renamo's Conduct in the War

Refugee reports suggest three typical types of Renamo operations. Within Mozambique, no geographical area will fit perfectly into the nominal description provided for each. But this appears to be the best way to disaggregate three perceived types of behavior in order to explain them. The three types of areas refugees described are (1) tax areas, (2) control areas and (3) destruction areas.

Tax Areas

Tax areas tend to be rural areas in which the population is extremely dispersed. Each family lives on the land it is farming or in small, extended family hamlets. Renamo combatants move freely through such areas and routinely (weekly or monthly) visit the farmers. They demand a contribution of prepared and/or dry food (food grain or flour), chickens and goats, perhaps some clothes, a radio or other possession. They demand at will young girls or married women for sex.

To the degree that it is necessary to transport the resources that Renamo extracts, the local people are obliged to serve as porters. (Renamo appears to have virtually no mechanized transport anywhere in Mozambique.) Their trips tend to be of short duration (counted in hours rather than days). As a general rule the porters are permitted to return to their homes when their service is completed. The journeys are not as harsh as those conducted in the other types of areas; reports of beatings of porters tend to be exceptions. If the family refuses to submit to these demands, they are likely to be severely from refugee reports that local people are not permitted to flee the area. But the burdens placed upon them are more moderate than those placed on the population of other areas. Renamo's policy appears to be to maintain a level of taxation and abuse insufficient to motivate the population to risk the security dangers, economic perils and social disruption represented by escape.

Murder of a civilian by Renamo in such areas is the exception rather than the rule. Nonetheless, Renamo combatants are reported to target individuals possessing resources whose origin, in the minds of these combatants, links them to government forces. The possession of new clothes, a radio or any type of army-issue apparel such as a belt or cap - perhaps even a bag of salt - may be enough to trigger torture or death. Such incidents, however, appear to be sporadic.

This sort of administration does not appear very different from the conduct of many guerrilla organizations. What is somewhat unusual, however, is the reported absence of effort to explain to "the people" the purpose of the insurgency or to provide positive services - security, administration, schooling, health care - to the local farmers. The relationship between Renamo and the population appears to revolve solely around the extraction of resources - strictly by force, without explanation, with no tolerance for refusal and without reciprocation.

Control Areas

About one-fifth of the refugees interviewed reported that they had lived in areas under the prevalent control of Renamo combatants. Some of these refugees had resided there for many months, others for many years. A significant number provided detailed descriptions of the manner in which these areas are administered.

According to the refugee reports, two principal types of civilian populations reside in these areas: indigenous local populations, referred to in Portuguese as naturales, and a population that was abducted from other areas and involuntarily marched into the control areas. According to the refugees, both of these populations are captive, detained against their will and prohibited from leaving.

Each of the control areas described by the refugees appeared to have a somewhat different system of organization. For purposes of explaining the general thrust of their reports, a control area tends to be nominally divided into three subcategories: (A) combatant bases, (B) field areas and (C) dependent areas.

(A) Combatant Bases

Combatant bases were of two types: those reserved principally for permanent, resident combatants; and those reserved principally for combatants who are passing through an area or who are, for whatever reason, temporarily stationed there. Both areas are managed in roughly the same fashion. They are served by a staff of exclusively male captives who provide food, water, cleaning and other support services. Women are provided on demand from other sectors of the control areas.

Refugees who resided in control areas, as well as other refugees who had contact with Renamo, said that its combatants represented indigenous language groups from all major regions of Mozambique. Forced recruitment was said to be the principal method through which these men had been impressed into service. Two refugees provided detailed accounts of their own escapes from apparent forced recruitment roundups. One said that his group, awaiting disposition, had been locked in a guarded house without food and water for such a long time that four of the men had starved to death. A few refugees said that they witnessed either the recruitment of young children (age 10 and over) or were themselves the victims of indiscriminate shootings or beatings by such young Rename combatants.

(B) Field Areas

Second are permanent agricultural lands, significant expanses of Renamo farms or plantations. The workers on these farms are captives who toil at a regular schedule, usually long hours during a six-day week. They do not benefit from the production of these fields. Their work is closely supervised, and physical punishment, in the form of beatings, is used to motivate those who take unauthorized rest or who refuse to continue. The worker population of these areas is reported to be principally older children of both sexes and adult males and females. This age group is also used as porters. Unlike those in the tax areas, the marches undertaken by these porters tend to be of longer duration - sometimes a week or more roundtrip. For some the portering is occasional, but many of the porters appear to perform this as a fulltime service, making continuous rounds of arduous trips.

The refugees report that these trips usually begin at dawn and continue until evening-perhaps 6 AM-8 PM. More recent arrivals report that, increasingly, the porters are not permitted food on these trips and are allowed to drink water only when they pass streams. Some report that they are often forbidden to drink because some make their escape during these rests. The loads are heavy; the pace, quick. The porters carry food toward the larger bases, and meat from some wildlife areas back to the bases in their home control areas. They also carry weapons, ammunition and the personal supplies and equipment of military units as they are shifted over long distances between provinces. None of the roughly 20 percent of the refugee sample that served as porters said they made trips to the sea. When they reached the destination bases - wherever these may be located - the porters deposited their loads outside. According to those who have served as porters, Renamo combatants then pick up the materials and take them into the bases.

Porters reported uniformly that discipline on these forced marches is extremely harsh. Those who cannot keep up, who rest without authorization, drop their burdens or refuse to continue are routinely beaten very severely, sometimes until they die. Over half of those who served as porters said they had witnessed cases of porters beaten to death or executed for these reasons.

When the captives are not being used as porters and the amount of such time varies, depending on whether they are full- or part-time porters - they are usually returned to Renamo-controlled fields. Some areas, which are not organized into large Renamo controlled fields, permit the captives to have their own land, but extract a great proportion of the harvest, leaving only a small amount for the family.

Another function of the young girls and adult women is to provide sex to the combatants. Refugee reports indicate that these women are required to submit to sexual demands, in effect to be raped, on a frequent, sustained basis. The rape may occur in field area residences. According to those who have served as porters, women are frequently raped along the transport routes. Women are also dispatched from the field areas to combatant bases on demand. One of the frequent refugee complaints (verified by medical relief workers in some of the refugee camps) is the level of venereal disease proliferated by this practice. Young girls and women who resist sexual demands are beaten severely, a punishment that might also be inflicted on the husband or father of the woman who resists. Such punishment reportedly can, at times, include execution.

(C) Dependent Areas

Dependent areas are populated by the elderly, who cannot serve in other, more taxing functions, and by young children. Refugee reports indicate that the elderly women are responsible for the care of the young children, and the elderly men are responsible for collecting thatch and building houses for the combatants. (In some areas, elderly men and women are also used as field workers and porters. Some of the more brutal accounts of beating and killing of porters who could not continue involved the elderly as victims.) Although specialization of function by age and sex is sometimes reported, only some of the control areas are reported to split families and oblige the members to live near their respective locus of service function.

Survival of Captives in Control Areas The captive population is reportedly not materially remunerated, nor are they provided with in-kind resources from their productive effort. Instead, in their spare time, the field workers are permitted to grow food for themselves. Foremost, their "spare time" consists of one day per week; during the other six days, they labor in the fields from early morning until dark, and the full-time porters have particularly disruptive schedules.

When people work on the alternative system - the highly taxed, individual production system - drought appears to increase the levy. In areas in which food is extremely scarce (and this number is not insignificant), the captive laborers, on their own initiative, gather wild mushrooms, leaves and berries, all natural survival foods. Some refugees reported that even these meager foods have been confiscated by the Renamo administration. At times, Renamo combatants or police dispatch captives into the bush to gather these survival foods. Several of those interviewed had escaped during these unsupervised forays.

Increasingly, those who served as porters testified that they were forbidden to eat during the trips. They alleged that some people died from lack of food. The Renamo guards and police who oversee these transport columns eat regularly and, some say, carry canteens. Some porters suggested that they were deprived of food and water in order to keep them too weak to attempt escape.

Clearly, however inadequate the food supply, Renamo appropriates most of the food produced for the organization and its combatants. When food is short, they always eat first. The refugees' allegations of food deprivation explain the wasted physical condition of the thousands of men, women and children arriving in asylum countries naked and severely malnourished.

Control Area Security and Surveillance Given the unusually high rate of resource and energy extraction from the captives in the control areas, it is curious that more have not fled and that those who have fled tend to arrive in such pitiful physical condition. The explanation, according to the refugees who have resided in these areas and later escaped, lies in Renamo's effective security and surveillance system.

The Renamo police, called majuba or mujiba, are at the heart of the system. According to the refugees, the police fall into two categories: (1) former tax collectors and petty officials (some called regulos) associated with the previous Portuguese colonial administration, who were replaced by the Frelimo government after independence from Portugal; and a small number of previous Frelimo members who became disaffected; and (2) captives identified by Renamo and obliged to serve in this role. These captives may be motivated on the one hand by punishment should they either refuse the role or fail to effectively implement it, and on the other with some degree of preferential treatment. In general, the refugees said that these police tend to be armed with cutting instruments (machetes, knives, bayonets, axes) rather than with firearms.

The police are deployed in two ways. Some are stationed along the geographical perimeters of the control areas, which are usually marked by physical features such as rivers and mountains or by a circle of combatant bases, which lend themselves to police patrolling. A second set of police reside in the captive areas and supervise the fields, the residence areas and the movements of the captives within the control area.

Attempted escape from a control area is often treated as a capital crime. One man reported that because he attempted to escape, he was put in a Renamo jail. The majority of those who resided in a control area report that they witnessed severe punishment or death inflicted on captives apprehended while attempting to escape. Wives and children left behind by those who made successful escapes are sometimes executed in retribution. Refugees report that, in general, the Renamo executioners used these punishments to discourage others; executions of captured escapees take place in front of all the captives from a given area gathered together in a central place. One refugee reported that during a portering job he was led to a small clearing and shown the decomposed bodies of many men, women and children. Renamo combatants told him that those apprehended while attempting to escape were taken here and executed. The refugees almost uniformly reported that these demonstrations and their own contextual understanding of Renamo's methods had a chilling effect on those considering escape.

It appears that not until death becomes a real possibility in their minds do the captives consider risking the dangerous escape attempt. But they explain that a number of additional constraints encumber such actions. First, both naturales and abducted captives find that if they have families with them, escape is more difficult. They state that a larger group is more likely to be apprehended; on-the-spot execution is, in their minds, a likely outcome. Second, captives who are natives of other areas and were abducted to control areas were unfamiliar with the new terrain, uncertain of bush routes they had to follow to avoid detection.

Finally, several refugees reported that an increasing constraint on escape is that a part of the captive population is too starved, too weak, too old or too sick to undertake flight. Some parents might send those children who are sufficiently grown and still healthy enough to make the journey, perhaps with an adult natural as a guide. The escapees almost invariably travel at night, surviving on wild fruits and leaves and walking through the bush, never on the roads. (In one recent incident, however, captives said they were ordered by Renamo combatants to leave one control area and their houses were burned.)

Renamo appears to have little difficulty in replacing individuals who have been killed, have died of starvation, have escaped or are too weak to produce. The principal source of new captives appears to be abductees taken in "destruction operations" described later; occasionally, Renamo appears to abduct individuals from tax areas.

Control areas are distinguished from tax areas by the higher degree of extraction of energy/labor/food production from the population. The refugees describe this rate of extraction as unbearable; the wasted physical condition of escapees who reach safety provides physical evidence for this assertion. As in tax areas, this extraction appears to be the only type of relationship between Renamo and these captives. There were virtually no reports of Renamo providing protection, remuneration, in-kind assistance, administration (except for security and surveillance), social services or material or any other type of perceived reciprocity. Renamo makes almost no reported effort to explain the purpose of its efforts or the nature of its goals, or to enlist the loyalty-or even neutrality - of the population. The only reciprocity the captives appear to receive or to expect is the opportunity to remain alive.

A few refugees who lived in control areas reported that their villages had schools for their children. One of these refugees stated that his children were taught to inform the Renamo police if they spotted strangers walking in their area and were also taught to read and write. Beginning with ages 10 and 11, however, he reported that the children are used for other purposes: some are old enough to work in the fields; others are recruited to become Renamo soldiers. One refugee reported that his village had a school and that attendance was obligatory; he added, however, that the main school activities were working in the fields and building houses and latrines. (One refugee reported that he was originally captured by three Renamo soldiers, whom he described as being 10-12 years old and particularly vicious.) One of two refugees who lived in an area with a school also reported that the town had a health clinic, although it contained no medicines.

Destruction Areas

Areas not designated as tax or control areas might fall into a third category-areas targeted for destruction. Such areas often include "villages" of 40 or more families residing in a small hamlet - although these also include villages of several hundred families. Villages designated as "settlements," especially those developed by the Mozambican government for refugees who have returned from neighboring countries, appear in some cases to be priority targets. Entire geographical areas might be subject to the destruction zone designation; sometimes only the larger villages within such areas are targeted for destruction. Refugees from various areas of Mozambique reported consistent patterns and methods of attack, with similar consequences. Not all villages in Mozambique have been subject to such attack, but from the geographical diversity of the refugees reporting, it appears that significant areas have been the object of this type of activity, apparently aimed principally at the destruction of these population centers.

The destruction program can proceed through up to three nominal stages. First, some villages targeted for a destruction effort appear to become the subject of careful, advance intelligence collection. Renamo combatants reportedly gather intelligence by visiting local farmers in their fields - or by abducting one or two villagers, who are subjected to a debriefing and sometimes ordered not to return to the village. The questions pertain to the deployment, if any, of government soldiers and/or militia in the village; the nature of their arms; the identity of civilian officials of the government - the identity of the village chairman, secretary and other prominent citizens and the location of their homes; the identity of individuals owning valuable possessions - new clothes, a radio and so on; and the habits of the population.

Second, depending on the area, Renamo combatants may follow up with a "political" visit (although in some cases this second stage is combined with the attack stage). During this stage, Renamo combatants might proceed directly to the homes of government officials and apprehend them and in some instances their wives, children and other relatives who are in the home. Refugees report that such persons are usually executed, sometimes after the village is assembled to witness the incident. One refugee reported that a group of officials and their wives and children were burned alive in their homes after the husband had been immobilized with an ax wound, while villagers were forced to watch. During this stage, Renamo combatants sometimes instruct the people to disperse from their village to their individual fields, tax areas. They may also be advised to go to control areas.

The third stage for villages and areas designated for destruction is the actual military attack. Such attacks were generally said to occur in the early morning hours, just before or after dawn. In some cases the attack group appeared to move from village to village in a selected area, conducting the same operation in each. The attacks seemed to be organized and executed in the same manner, regardless of whether the village was defended by government soldiers or militia. (As a general rule, civilians possess no firearms.) Thus, if no government soldiers or militia are present, the attack is conducted against unarmed civilians. In most of the small villages, a handful of government soldiers or militia were reported to be present. According to the refugees, these soldiers would often be rapidly overwhelmed by the more numerous attacking force and would either be killed or forced to flee soon after the attack began. Larger villages sometimes have larger defending forces, and in these there is a longer, and sometimes successful, resistance.

The attack stage was sometimes reported to begin with what appeared to be the indiscriminate firing of automatic weapons by a substantial force of attacking Renamo combatants. Because the attack begins during the very early morning hours, this force usually has the element of surprise. In some villages, the firing is sometimes reported to be directed into the houses of the civilian population, at inhabitants who attempt to flee and at the handful of government soldiers or militia, if any, who return the fire. The government soldiers reportedly aim their defensive fire at the attackers; the Renamo forces shoot indiscriminately into the village. Some refugees perceived that the attacking force had divided into three detachments: one conducts the military attack; another loots homes of valuables, mainly clothing, radios, food, pots and other possessions; a third moves through the looted houses, setting fire to them with pieces of burning thatch. There were several reports of schools and health clinics being typical targets for destruction. The destruction of the village as a viable entity appears to be the main objective of such attacks.

This type of attack causes several types of civilian casualties. As is expected in guerrilla warfare, some civilians are killed in crossfire between the two opposing forces, although the refugees say that this tends to account for only a minority of the deaths. A larger number of civilians in these attacks and in other contexts were reported to be victims of purposeful shooting deaths and executions; of axing, knifing, bayoneting, burning to death, forced drowning and asphyxiation; and of other forms of murder in areas which have no meaningful resistance or defense. Eyewitness accounts indicate that when civilians are killed in these indiscriminate attacks, whether against defended or undefended villages, children, often together with mothers and elderly people, are also killed.

Varying numbers of civilian victims in each attack were reported to be rounded up and abducted, apparently to control areas. During the initial abductions, they are sometimes simultaneously used as porters to carry away goods confiscated in the attacks on their villages. From refugee reports, treatment of these porters seems to be consistent with that of the regular porters described in the preceding section on control areas.

Deaths were also reported in instances when Rename combatants, in the absence of resistance or defense, were reported to chase the villagers toward natural barriers, such as rivers, or otherwise use the physical surroundings to cause their deaths by drowning or asphyxiation.

The survivors of attacks who are not severely wounded and who are not abducted appear to face three destinations: some flee to their fields, including the tax areas; some flee to government-controlled towns, typically district capitals; and some flee to neighboring countries.

Refugees residing in areas apparently marked for destruction were asked repeatedly why they did not, upon learning of initial abductions and particularly after the (second stage) "political" visit, flee their villages. Many responded that they did not realize, at each of the early stages, what fate awaited the village; they had been reluctant to leave their home villages until there was absolutely no other choice. Some of the villages had been subjected to several Renamo "third stage" attacks before the final successful attack which drove the villagers from their homes. In each of the earlier attacks, people were killed. But it sometimes appears that until someone in the immediate family was killed or home and property destroyed, they were most reluctant to abandon the homes and land with which they seemed to identify very closely. Moreover, some expressed fear of the dangerous journey to safety and uncertainty about the safety of the alternative destination area. Refugees also mentioned economic survival: how would they eat and where would they get food if they left their villages?

In both tax and destruction areas, a small number of cases were reported of "targeted retribution" against small children. These involved, for example, the mutilation and subsequent killing of young children in retaliation against parents who fled a Renamo visit or attack. Most civilians, however, took their children with them when they fled; the child retribution cases appeared to affect mainly the small number of families who could not find a child at the moment of flight. Mutilation by Renamo combatants or police of adults - the severing with a knife of mainly ears, but also lips, noses and limbs - was reported by a small fraction of refugees.

The foregoing description does not encompass every type of Renamo operation in Mozambique, but it does characterize in nominal categories both its general types of conduct and its relationship with the respective civilian populations reported by the great majority of the refugees. The purpose of dividing these accounts into three nominal types of areas (tax, control and destruction) is to convey complex reports in a more understandable form. No area conforms strictly to one pattern alone.

The great majority of the 196 interviewed refugees whose migration was caused by conflict cited Renamo actions as the reason for their flight. Roughly 40 percent reported witnessing the murder of civilians principally by Renamo combatants and Renamo police in the absence of resistance or defense. The 69 refugees who arrived at their current locations in 1987-1988 reported roughly 600 such murders. The refugees provided eyewitness or other credible accounts about these killings, which included shooting executions; knife, axe or bayonet killings; burning alive; beatings; forced asphyxiation; forced starvation; forced drownings; and random shooting at civilians in villages during attacks.

The shootings listed above do not include about 200 persons killed in crossfire between Renamo and Frelimo government soldiers, which this report attributes to neither side in the conflict. The approximately 600 murders attributed to Renamo by the 1987-1988 arrivals represented about 94 percent of such murders reported by these refugees (of the remainder, about 3 percent were attributed to Frelimo and about 3 percent were attributed to "unknown" parties). An additional 10 persons who drowned in the Zambezi River by accident while escaping from Renamo pursuit are not counted among the roughly 600.

About 50 of the 600 reported Renamo murders were reported to be of young children. Of these, about 30 fell into the category of murder for retribution, as described in the preceding section, and included the children of Frelimo civilian village officials killed in the second stage of the campaign against the destruction areas.

The results in the other abuse categories were similar. Nearly 40 percent of the refugees had direct knowledge of Renamo's forced portering on the civilian population; about 20 percent had themselves served as porters. Over 70 percent of those who had served as porters witnessed severe beatings of porters who could not keep pace, stumbled or dropped their loads. About 60 percent of those who served as porters reported that they had seen or had credible reports of captive porters who were beaten to death or executed along the route for these same reasons. Four of the refugees who had served as porters described incidents in which they themselves were severely beaten. No portering complaints against Frelimo government soldiers were received.

More than 15 percent of the refugees reported patterns of systematic rape of civilian women by Renamo combatants. One rape complaint against government soldiers was reported. About one-third of those reporting house burnings saw their own houses burned. No complaints of house burning were reported against government soldiers.

Nearly 40 percent of the refugees complained about abduction of civilians, a pattern most typical of the nominal destruction areas. Of these, over a quarter had themselves been victims of abduction. Ninety-four percent of these identified Renamo combatants as the abductors; the remainder were split between government soldiers and unknown parties.

Looting and forced contribution of resources was reported by nearly 60 percent of the refugees. Of these, some 45 percent reported that beatings were used when necessary to extract resources. Of the looters, 93 percent were identified as Renamo combatants, 6 percent as government soldiers and one percent unknown.

Mutilations, reported by five percent of the refugees, were all attributed to Renamo combatants.

Some 21 percent of the refugees provided eyewitness accounts of life in control areas based on personal residence there as captives. About 60 percent of the refugees said they witnessed or had credible knowledge of the severe punishment or execution of captives who had been caught by Renamo attempting to escape from control areas and other activity areas.

For the 1987-1988 arrivals, discounting non-conflict migrants, 91 percent of the refugees' reports were categorized by the author as "very negative" toward Renamo; 5 percent as "negative"; one percent as positive. Three percent offered no report and no complaints.

Frelimo's Conduct in the War

Complaints by 1987-1988 refugees about government (Frelimo) combatant conduct were both serious and quite different from the reports of Renamo conduct. In general, refugees described the principal role of the Frelimo soldiers as defending their villages against the Renamo attackers. Many of the refugees reported that in attacks on small villages that might have had from two to 10 Frelimo soldiers and militia defending them, the government soldiers were quickly overwhelmed by a far larger insurgent force. In some cases, refugees said that Frelimo soldiers fired their weapons and attempted resistance just long enough to give civilians an opportunity to flee.

Accounts of serious abuses in individual incidents attributed to government soldiers were reported in credible detail. However, they were small in number by comparison to complaints concerning Renamo and did not appear to represent systematic problems of discipline or command and control. Because complaints were relatively few, the 1987-1988 data can be summarized as follows: compared to the 94 percent of murders attributed to Renamo, about 3 percent (roughly 20) were ascribed to Frelimo (3 percent were attributed to "unknown" parties). No reports of death by forced drowning; by axe, knife or bayonet; by burning alive; by beatings; by forced asphyxiation; or by random shootings were attributed to Frelimo soldiers by this sample. Compared to 96 percent of child murders attributed to Renamo, only 4 percent were ascribed to Frelimo. No complaints of forced porterage or burning of houses were ascribed to Frelimo. The 169 arrivals during the 1987-1988 period attributed one complaint of rape to a Frelimo soldier. Three cases of abduction by Frelimo soldiers (4 percent, versus 96 percent for Renamo) were reported. Six cases of looting or forced contributions (6 percent, versus 94 percent for Renamo) were reported. There was also one complaint that Frelimo government soldiers had destroyed a farmer's crop, and one complaint of a man who had been arrested because in earlier years he had resisted a Frelimo government order to move into a central village. No reports of mutilation were attributed to Frelimo soldiers.

In general, informed sources within Mozambique reported that the human rights conduct of Frelimo government soldiers has been on a markedly positive course since 1986; Renamo conduct, however, appears to have worsened.

In summary, after discounting the non-conflict related migrants for the entire sample, refugee reports on Frelimo and Renamo are characterized by the author as follows:

Very Somewhat No

Positive Negative Negative Complaint

Frelimo 11% 7% 10% 72%

Renamo 1% 91% 5% 3%

Thus, the combined "negatives" concerning Renamo were 96 percent, versus 17 percent for Frelimo. Combined "positives" plus "no complaints" concerning Frelimo were 83 percent, versus 4 percent for Renamo.

Pre-1987 Report Differences

The pre-1987 arrival sample was too small to provide a useful basis for comparison with the 1987-1988 arrivals. Because of the limited sample size and because a larger proportion - about one-third had migrated because of conditions not related to the conflict, it must be treated with some reserve. The attribution of human rights abuses to Renamo versus Frelimo among this group is roughly the same as for 1987-1988 arrivals. The proportion of non-murder abuses (rape, looting, forced porterage, mutilation and so on) attributed to Renamo is slightly higher. Of the witnessed murders, however, 68 percent were attributed to Renamo, 32 percent to Frelimo, none to "unknown." The proportions arising out of these particular numbers should be given only relative weight because of the limited size of the sample (particularly after non-conflict migrants are deducted) and the consequent possibility that the dimensions of a single incident could disproportionately affect the overall results.


From 48 districts in northern, central and southern Mozambique, in 25 refugee camps in five countries separated by as many as 1,500 miles, the nearly 200 Mozambican refugee accounts are strikingly similar. If that sample is reasonably representative, their accounts, corroborated in large measure by independent experience of some religious and relief assistance workers, point to certain unavoidable findings.First, the level of violence reported to be conducted by Renamo against the civilian population of rural Mozambique is extraordinarily high. Roughly 170 refugees, each representing one family, who arrived in 1987-1988, collectively reported about 600 Renamo murders of unarmed civilians in the absence of resistance or defense. (If the population estimates reported in the introduction to this report are correct, there are roughly 200,000-250,000 refugee and displaced families in Mozambique and in the neighboring countries, the majority of whom are conflict victims.) If the refugee reports are generally accurate and the sample reasonably representative, it is conservatively estimated that 100,000 civilians might have been murdered by Renamo in this manner.

The same 170 refugees report many hundreds of cases of systematic forced portering, beatings, rape, looting, burning of villages, abductions and mutilations. These patterns of systematic abuse represent many hundreds, if not thousands, of individual instances reported by this small sample. Conservative projections based on this data would yield extremely high levels of abuse.

The accounts by refugees who have fled from northern, central and southern Mozambique are so strikingly similar that they suggest that the violence is systematic and coordinated and not a series of spontaneous, isolated incidents by undisciplined combatants.

Second, the relationship between Renamo and the civilian population, according to the refugee accounts, revolves almost exclusively around a harsh extraction of labor and food. If these reports are accurate, the only reciprocity provided by Renamo for the civilians' efforts appears to be the possibility of remaining alive. Virtually no reports exist of attempts to win the loyalty - or even neutrality - of the villagers. The refugees report virtually no effort by Renamo to explain the purpose of the insurgency, its proposed program or its aspirations. If a significant sector of the population is in fact sympathetic to this organization, it was not reflected in the refugee accounts.

Third, serious complaints exist about abuses by some Frelimo government soldiers. But in both the murder and non-murder categories, only 3-4 percent of the complaints were attributed to Frelimo soldiers. They tended to be isolated reports, often from areas of the country most remote from Maputo. It appeared that there is a sustained trend toward improvement overall.

Fourth, the refugees and most independent sources rejected the assertion that much of the violence in Mozambique is attributable to neither Frelimo nor Renamo but instead to armed, unaffiliated bandits. It appears from this field research that violence by "freelance bandits" does not account for more than occasional, isolated instances of the high level of reported violence.



Ressano Garcia

Caia (2 subsites)

Casa Banana


A Luta Continua

Moatize (2 subsites)



Estima (2 subsites)


South Africa






Mazoe Bridge

















ra Bassa

































































Article copyright Cultural Survival, Inc.

CSQ Issue: