Citation
Journal of Agrarian Change, Vol. 11 No. 3, July 2011, pp. 396–419.
The 1990–92 Massacres in Rwanda:
A Case of Spatial and Social Engineering?
PHILIP VERWIMP
Until now, two main sets of arguments have dominated the debate on the nature of the
massacres that were perpetrated in Rwanda before the 1994 genocide. The first one
maintains that they constituted a response to prior attacks by the RPF, implying that they
should be regarded as military operations, rather than as acts of ethnic cleansing.The second
common line of argument is that these massacres served as pilot runs for the subsequent
genocide, implying that they were part of a plan that was not to see its full implementation
until 1994.This paper puts forth a third, alternative interpretation of these massacres.The
first of the aforementioned arguments, it is contended, does not take into account the detailed
evidence that is available on the killings: the fact that they took place in the context of the
civil war accounts for the timing of the massacres, but not for their genocidal character. In
turn, the second interpretation fails to situate these massacres against the agro-pastoral and
ideological background of the regime that committed them. By contrast, this paper shows that
the massacres took place in areas characterized by a specific history of spatial and social
engineering. They are best understood against the background of the processes of land
colonization, resettlement, depredation and dispossession of cattle and land that were under
way in the areas where the land was most scarce, and where the peasant society was being
subject to rationalization and remodelling from above.The paper concludes that pastoralism
was sentenced to disappear from Rwanda and that the massacres should be considered
instances of ethnic cleansing.
Keywords: massacres, pastoralism, state, social engineering, genocide, Rwanda
INTRODUCTION
Between 1990 and 1992, around 2,000 Tutsi civilians were killed in a number of massacres that
were perpetrated in Rwanda – mainly in the north-west of the country, but also in the
southeastern region of Bugesera. These massacres, while denied at the time by the local and
national authorities, have been well documented and are now considered a part of our body of
knowledge concerning the 1990–92 period of Rwandan history.Thanks to several high-profile
publications on human rights violations in Rwanda, there is no doubt that these massacres
indeed took place.1 According to reports by diplomats and human rights organizations, the
Philip Verwimp, Solvay Brussels School of Economics and Management and ECARES, Université Libre de
Bruxelles, Avenue Roosevelt 42, 1050 Brussels, Belgium. E-mail: philip.verwimp@ulb.ac.be
The author expresses his gratitude to the people who shared their knowledge on the subject over several years;
in particular, B. Kiernan, A. Desforges (†), A. Guichaoua, J.P. Platteau, T.Vandevelde and P. Uvin. Only the author
is responsible for the views expressed in this paper.The guest editors, three anonymous reviewers as well as seminar
participants at Yale University’s Order, Conflict and Violence Program have contributed substantially to improving
the paper. The author thanks Francois Bart for the permission to reproduce three figures.
1
The four most cited reports on human rights violations that focus on, and were published during, this period
are: (1) International Commission on Human Rights Violations in Rwanda since October 1990 (FIDH, March
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 397
anti-Tutsi pogroms of 1990, 1991 and 1992 were in fact organized by the national and local
authorities. These pogroms, which occurred in several different communes, took place in
locations that had been carefully chosen by the national leaders. The leadership mobilized the
Hutu peasants by spreading rumours (fabricated stories) in order to install fear and incite Hutu
hatred. One report, dated March 1993, has discussed the applicability of the term ‘genocide’ to
the killings that had already taken place prior to that date, while another, published in August
1993, argues that the killings comply with the international definition of genocide.
With regard to the interpretation of the massacres, particularly their cause(s), the scholarly
community is in disagreement. Generally speaking, two main sets of arguments have been put
forth in order to explain the massacres that were committed in Rwanda in the period between
1 October 1990 and 6 April 1994. The first one, formulated in slightly different versions
by René Lemarchand, Filip Reyntjens and Scott Straus (among others) maintains that these
massacres were linked to the war (Reyntjens 1994); constituted a rational response to attacks
by the Rwandan Patriotic Front (RPF) on the part of a population that felt threatened
(Lemarchand 2002); or constituted a response to RPF attacks by the government and the
local authorities in a context of war, insecurity and political uncertainty (Straus 2006).
All these authors have in common the fact that they place the war in the centre of their
explanations of the massacres – whether this is presented as having been exacerbated by fear on
the part of the population, in some versions, or driven by a response on the part of the leadership
with a view to re-establishing order, in other versions. Lemarchand adds that the response on the
part of the peasant population had a spontaneous character:‘They had no other choice but to kill
in order not to be killed.’ Straus applies this line of reasoning not only to the 1994 genocide and
the 1990–92 massacres, but also to those that were perpetrated in late December 1962 – at which
time between 5,000 and 10,000 Tutsi civilians were brutally put to death in the province of
Gikongoro. According to Straus (2006, 184–8), this latter massacre took place in reaction to an
attack on Kigali via Bugesera (in the province of Rural Kigali) by armed Tutsi.
The second set of arguments regards these massacres as trial runs in preparation for the 1994
genocide. In her seminal book on the Rwandan genocide, Alison Des Forges writes that ‘[t]o
execute a campaign against Tutsi effectively took practice. Before the grim background of war,
economic distress, violent political competition, insecurity and impunity, and to the accompaniment of virulent propaganda, radicals staged the practice for the catastrophe to come. The
rehearsals took place in more than a dozen communities’ (1999, 87).
With the benefit of hindsight, these massacres may seem like pilot runs or rehearsals leading
up to the 1994 genocide, but this latter event cannot be considered a satisfactory explanation
for events that preceded it. Rather, the explanation for the 1990–92 events needs to be sought
in facts and events that either took place at the same time as these massacres or preceded them.
Des Forges argues that the massacres were organized by Habyarimana and his supporters, and
adds that the regime used ethnic violence to its advantage. At a time when Habyarimana was
facing military and political threats (1999, 87–8), these massacres strengthened his position,
fostered Hutu solidarity and rallied the Hutu behind a united cause. This author’s analysis
portrays these massacres as having been perpetrated for instrumental reasons (in order to
1993), which implicates the country’s highest-level authorities in the organization of the killing of 2,000 Tutsi in
several locations throughout Rwanda; (2) the report published by the US Department of State in February 1993,
which describes the massacres of the Bagogwe (January 1991) and of the Tutsi in Bugesera (March 1992) (in March
1991, the US Department of State had already published a report on the January 1991 massacre); (3) two reports
published by the Rwandan human rights group ADL in December 1992 and December 1993, respectively, which
describe in detail several massacres and instances of human rights violations against the Tutsi; and (4) the report by
the UN special rapporteur on Rwanda that was released in August 1993, which maintains that these massacres
comply with the international definition of genocide.
© 2011 Blackwell Publishing Ltd
398 Philip Verwimp
consolidate power, heighten ethnic tensions and polarize society), which is in stark contrast to
the proponents of the ‘war argument’, for whom the massacres were a desperate expedient, a
price that had to be paid and a response to prior attacks that had disturbed peace and order.
In order to understand why it was that the regime organized attacks on Tutsi citizens from
the very beginning of the civil war, it is necessary to look at the details of the massacres that
were perpetrated prior to 1994. For this, it is useful to draw on the report published in March
1993 by the International Commission of Human Rights Investigators – a group of experts
who investigated several of the massacres that took place in Rwanda in January of that year.This
report provides a wealth of detail on these massacres and, in this paper, special attention is given
to the sequence of events in each of these early massacres.
It is probably futile to look for one single objective or explanation behind the massacre policy
of 1990–92. It is highly likely that several different factors played significant roles at the time
– including the threat posed by the RPF and the governments’ wish to show its resolve,
consolidate its power and foster Hutu solidarity. In view of the arguments put forth in the
literature, however, one major factor is arguably missing in the discussion: the peasant ideology
professed and practiced by the Habyarimana regime. Taking the latter into account allows for
an alternative explanation of the massacres, or at least for a richer interpretation of the
motivation behind the massacres.This third interpretation is as follows: the Habyarimana regime
had adopted a policy of agricultural extensification – turning all available land, such as pastures,
marshes and forests, into cultivable land – as opposed to a policy of intensification. In Rwanda,
this policy came up against the land frontier in the late 1980s; in other words, all the land had
by then been converted to agricultural activity. In their turn, pastoralist groups such as the Hima
and the Bagogwe used land as pasture for their cattle, living off the cattle itself and the trade
in meat and dairy products. They did not cultivate and were therefore considered a nonagricultural group. Under the predominant ideology of the Second Republic, which portrayed
itself as a Peasant-State, pastoral groups were marginalized, and pastoral lands were converted
into land for cultivation and into paysannats – the prime agrarian settlement scheme. Pastoralism
as a way of life did not fit within the agrarian order of the Second Republic, which was built
on a vision of hard-working smallholder peasants.2 The regime thus used the opportunities
provided by the civil war in order to claim the last remaining parcels of land by finishing off
the last remnants of pastoralism in Rwanda.Therefore, these massacres can be described as a case
of ethnic cleansing. The point was not that the Hima, the Bagogwe and the Tutsi owned cattle
(for many rich Hutu and the dignitaries of the Habyarimana regime did so, too), but that their
pastoralist livelihood did not fit in the Peasant-State. One can romanticize peasant cultivation,
but when such an ideology is combined with racism towards ethnic groups regarded as
non-peasant, such as the Tutsi, it can assume an extremely violent character.3 This interpretation
is supported by the socio-economic geography of the massacres, and for that reason this paper
seeks to draw attention to the spatial dimensions of violent conflict.4
A.M. Brandstetter is one of the few scholars who have analysed the peasant ideology of the
regime. In her 2001 article on purity and violence in Rwanda, she puts forth the argument that
the 1994 genocide was an act of purification of the body politic whereby the sons of the soil (the
peasants) sought to clear the bush. Although this author did not address the 1990–92 massacres,
and nor did she examine the agrarian resettlement schemes, my own analysis has many aspects in
common with hers. For example, she argues that ‘the genocide, through its violence, was meant
2
For an institutionalist and political economy perspective to conflict between agriculture and pastoralism,
see Platteau (2000) and Salih et al. (2001).
3
For a discussion of the link between romanticism and violence, see Kiernan (2007).
4
See, for example, Raleigh et al. (2010); also see Nathan (2005).
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 399
to implement the exclusion of the pastoralists from the project to constitute an agrarian society’
(Brandstetter 2001, 68; author’s translation from the German text). In their turn, Bézy (1990) and
Newbury (1992), writing on rural development in Rwanda, have highlighted the limitations of
the extensive land policy: agricultural production was increased only because more land was taken
into cultivation, not due to agricultural innovations or intensification. By the end of the 1980s,
the physical land frontier had been reached: there was no more land available to be taken into
cultivation (Bézy 1990; Newbury 1992).
CONCEPTUAL FRAMEWORK
Rwanda’s history can be traced through the management and settlement of land, driven by
political power on the one hand and population growth on the other. As long as Tutsi held
political power, they earmarked land for pasture. When that power disappeared, the land was
turned into agricultural land. Habyarimana was not against cattle – presumably he recognized
the value of having a cow for dung as fertilizer. The point made by Bart (1993, 185) is that he
disfavoured the traditional way of holding cattle, meaning the herding of a large number of
cattle that grazed on pastures. For him, this was an inefficient use of the land.
The reorganization of agrarian space, land settlement in the paysannats (organized settlement
schemes), umuganda (mandatory communal labour), the development of the five-year plans as
well as the nationwide anti-erosion campaign are part of the agrarian developmental state that
Rwanda became under the Second Republic. The developmental state was organized much in
the same way as mapped out by James Scott in his 1998 book Seeing Like a State. Scott
characterizes four elements constituting disastrous social engineering: the administrative ordering of nature and society, or its ‘legibility’; a high-modernist ideology; an authoritarian state; and
a prostrate civil society. Ethnic cleansing may be considered as an especially extreme form of
social engineering.5 This occurs when the social engineers not only regard space, land, cattle
and settlements as malleable factors, but the size and the composition of the population itself.
Scott’s insights apply clearly to the case of Rwanda.
The Administrative Ordering of Nature and Society or its ‘Legibility’
Rwanda was a highly and tightly organized society. The penetration of its administration into
the hills was unmatched in Africa (Guichaoua 1997; McDoom 2009). From the ethnic identity
card system, the detailed registration of births, marriages and deaths at the communal level, the
policy of ‘ethnic equilibrium’ in schools and in the administration, the parallel organization of
the party and the state from the national down to the cellular level, to the policy to keep people
in the rural areas, the Second Republic was neatly organized, as a pyramid from the top to the
bottom. Many observers were stunned by the degree of organization and thus by the presence
of statecraft in the life of ordinary Rwandans. The desire to order was not limited to its
inhabitants, but also applied to nature, as witnessed in the drive to reorganize agrarian space,
land settlement and agrarian order all together.
High-Modernist Ideology
This is defined by Scott as a ‘muscle-bound version of the self-confidence about scientific and
technical progress, the expansion of production, . . . the mastery of nature . . . and, above all, the
5
See Mann (2005) for a treatment of the common ethnic cleansing roots of many advanced democracies and
Mazower (1998) on mass population movements and forms of ethnic cleansing in Europe in the twentieth century.
© 2011 Blackwell Publishing Ltd
400 Philip Verwimp
rational design of social order’ (1998, 4). It must not be confused with scientific practice. What
is meant here is a coherent set of beliefs that are not open to question. The design of five-year
production plans, the mobilization of the entire peasantry in weekly umuganda, the expansion
of the paysannats, the denial of crop failure and famine conditions in the south in 1989 and the
nationwide anti-erosion campaign testify to the revolutionary beliefs held by the leaders of the
Second Republic. On the latter campaign, Guichaoua (1991) writes that it illustrates how a
standardized, agrarian order, implemented with geometric precision, is unable to accommodate
variation in soil quality, steepness and local needs, and as such invites peasant resistance
(Guichaoua 1991). While communist countries adopted industrialization as their version of
the developmental state, the Agrarian Nation that Rwanda would remain under the Second
Republic, founded on a ruralist and peasant ideology, was Habyarimana’s version of the
developmental state.
An Authoritarian State
The authoritarian character of the Second Republic is demonstrated by several of its features.
Chief amongst these, we find the centralization of power (military, executive, party) in the hands
of the president; the carrot-and-stick policy employed in the coffee sector (Little and Horowitz
1987, 1988; Verwimp 2003); the fact that all Rwandans were required by law to be members
of the MRND (Mouvement Révolutionaire Nationale pour le Dévéloppement); the prohibition from forming other political parties; the submission of the judiciary to the authority of the
single party; the weekly animation sessions in honour of the president; the forced removal of
people from their land in order to create tea plantations; the organization of mock elections;
and the killing of political adversaries (in 1976 and 1988).6
A Prostrate Civil Society
This is defined by Scott as the lack of capacity to resist state plans. In his 1998 book Aiding
Violence, Peter Uvin paints a bleak picture of civil society in Rwanda. In Chapter 8, entitled
‘And Where was Civil Society?’, Uvin describes exactly what James Scott has in mind: a weak
and usurped civil society, incapable of making a fist when most needed. According to Prunier
(1995), the MRND was totalitarian. Its first letter ‘M’ stands for ‘Movement’, and the party
manifest said that its task was to mobilize all living forces for the benefit of the nation. Hence
there was no need for organizations outside the party. Even the highest religious authority; the
Archbishop of Kigali, was a member of the Central Committee of the MRND until ordered
to resign from that position by the Pope.
POPULATION, SPACE AND SETTLEMENTS
The strength of the state was used to remodel agrarian space, register and control the
population, and replace politics by development. But it did not stop there. Social planners
considered the size and the composition of the population as malleable factors, both at the local
level and the national level. This is best illustrated by the rural-to-rural migration and resettlement into paysannat settlement schemes.
The Habyarimana regime promoted internal rural-to-rural migration and resettlement from
densely populated to less densely populated areas. In this way, newly colonized land as well as
6
Animation sessions took place once a week after umuganda and consisted of singing and dancing in honour of
President J. Habyarimana.
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 401
land previously earmarked for pasture was transformed into agricultural land. Olson, who
studied migration patterns in Rwanda, writes that:
After independence, increasing population pressure resulted in changing economic circumstances, such as rapid decline in farm sizes and available land per person. One
response was a high rate of out-migration from the areas experiencing the most pressure.
The destination of these migrants was influenced by political factors; the government was
interested in settling land previously used for pastoral activities so it promoted organized
settlement schemes in the East. (1990, 150)
A specific agricultural settlement scheme, the paysannat, was part of this resettlement plan.
Farm households were given a plot of land that they could cultivate on condition that part of
the plot, specified in a contract, was allocated to an export crop such as coffee or pyrethrum.
The latter was grown in the province of Ruhengeri on land in Mukingo commune previously
used as pasture by the Bagogwe (see below). The contract stipulated that division, fractionalization or renting out of the plot was not allowed (Article 4); and that upon signing the
contract, and at the latest six months after, the holder would renounce his rights to previously
held land and pastures (Bart 1993, 406; author’s translation).While the farmers in the paysannats
were in general better equipped (water tanks, pipes, mills, silos) and monitored (one agricultural
monitor per 120 instead of 750 households), this kind of contract was not compatible with
the traditional way of living of Rwanda’s farmers – in particular, in terms of marriage and
inheritance. As a consequence, Bart found married couples still living with their parents. He also
found many absentee owners (traders, army officers, civil servants) benefiting from the settlement scheme without residing in it.7
The First and Second Republics vastly expanded the area devoted to the paysannats and the
people living in them, allowing Bart (1993, 391) to write in 1993 that 1 in 20 farmers now
lived in such a settlement scheme. The authorities used the scheme to put in place a coherent
policy of land colonization and control of internal migration. At the regional level, the
paysannats dominated the landscape and the communes in which they were implemented.
During the first ten years of Habyarimana’s presidency (1973–83), the number of households
living in paysannats increased from 30,000 to 54,000 (250,000 people, see Figure 1; see also Bart
1993, 393), with the bulk of the increase in the first five years and with a new paysannat in
Mutara (Byumba province). Jean-Claude Willame (1995, 136) writes that the authorities
associated the paysannat schemes and the projects for integrated rural development with the
Hutu Revolution. This was a short cut from the side of the authorities, because the paysannat
settlement schemes already existed before independence and were a product of the colonial
regime. It does show, however, how much the First and Second Republics continued these
schemes and considered them their policy.
The population that came to live in the paysannats originated from the most densely
populated or poorest areas of Rwanda (Gisenyi and Ruhengeri in the north and Butare and
Gikongoro in the south). They left their ancestral rural lands to become modern farmers in a
new environment prepared and managed by the state. Figure 2 gives the example for the region
of Bugesera. François Bart (1993, 395), in his seminal book on the geography of Rwandan
agriculture, writes that farm households from densely populated areas in Rwanda came to live
in the paysannats, sometimes mixed with locals. All those residing in the paysannats, however, had
to adapt their way of living to the new settlement scheme. This meant that:
7
The Belgian administration for development co-operation assessed that the costs of the paysannats are out of
proportion with the benefits (Bart 1993, 405).
© 2011 Blackwell Publishing Ltd
402 Philip Verwimp
Figure 1 The growth (in per cent) in the number of people living in Rwanda and in the paysannats,
1973–83
80
33
Growth of
the general
population
Growth of
the population
in paysannats
In Masaka, in the paysannat of Icyanya; most of the 92 pastoralist households stayed and
accepted the new conditions thereby changing progressively their way of living. In the
paysannat in Ntongwe, in Mayaya, 11 out of 20 interviewed households declared to be
originating from the hill where from 1966 onwards the paysannat was established. They
said that the authorities obliged them to move their house alongside the new road. In this
case, the paysannat presents itself as an enterprise for the remodelling of the pre-existing
agrarian structure. Essentially, it proceeds with the colonization of new land more than the
reshaping of existing land structures. (Bart 1993, 395)
During fieldwork on the origins of the genocide in the province of Gitarama in 2004, we
found that the administration appointed an agronomist and an agricultural surveyor per
commune, as well as agricultural monitors (i.e. extension agents), who were assigned to three
cells at a time.8 Farmers who did not follow careful maintenance practices were punished with
a fine of 100 RWF per coffee tree.The agronomist and agricultural monitors came two to three
times a month to check on the coffee trees and to punish delinquent coffee growers. In the area
we visited, there used to be a paysannat where the agronomist and the monitors also organized
the weekly umuganda. Consequently, they had a lot of sway and power over the people in the
paysannat. The agricultural surveyor had a list of families to visit each season to collect data
about their fields and to find out how much each produced. This information was recorded to
keep track of the paysannat statistics and agricultural records.
8
A cell was the smallest administrative unit in Rwanda consisting of between 50 and 100 families. For an analysis
of the results of the fieldwork, see Pinchotti and Verwimp (2007).
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 403
Figure 2 Internal migration paths and the colonization of the Bugesera
Source: Bart (1993).
IDEOLOGY
The Peasant-State
At the birth of the Second Republic in October 1973, during a speech at the National
University of Rwanda, in Butare, the new President Habyarimana declared that ‘celui qui ne
veut pas travailler est nuisible à la société’ (translated: ‘the one who refuses to work is harmful to
society’). Habyarimana did not direct his speech against the radical leaders (professors and
students) who, as members of the Committee de Salut Public in the first half of 1973 had expelled
all Tutsi from the campus. On the contrary, he lamented against the feudal-monarchists:
The coup d’état that we did, was above all a moral coup d’état. And what we want, and
we would consider our action as failed if we do not reach this goal, what we want, is to
ban once and for all, the spirit of intrigue and feudal mentality. What we want is to give
back labour and individual yield its real value. Because, we say it again, the one who
refuses to work is harmful to society.9
9
Juvenal Habyarimana, in a speech on the occasion of the opening of the academic year in Butare, 14 October
1973.
© 2011 Blackwell Publishing Ltd
404 Philip Verwimp
This small extract contains the core of Habyarimana’s peasant ideology.10 In 1986, Habyarimana
said on two occasions that the peasants were the real employers of Rwanda, because they
allowed the State to function. On the occasion of the 25th anniversary of the existence of the
Rwandan Republic, on 1 July 1987, Habyarimana devoted his official speech to the glorification of the Rwandan peasant. He said that:
If in the 25 years of our independence Rwanda has known a lot of success in its struggle
for progress, if it has been able to take a number of important steps, it is in the first place
our farmers who made this happen . . . it is their total devotion to the work, every
day . . . their fabulous capacity to adapt, their pragmatism, their genius, their profound
knowledge of our eco-systems that allowed them to extract an amazing degree of
resources from their plots of land.
At the time, a commentator wrote that never before had such honour been given to the
Rwandan peasants (Ntamahungiro 1988). Four months later, on the occasion of the Government Council of 13 November 1987, Habyarimana ennobled the Rwandan peasant by
extending the term ‘peasant’ (Umuturage) to all Rwandans. The term umuturage was commonly
used in opposition to the civilized, educated, urban or bourgeois person. Umuturage was used in
a pejorative sense for the downtrodden, the uncivilized, the rural population. By using and
ennobling the term umuturage, Habyarimana wanted to invert the common meaning of the
terms. From now on, umuturage would be a noble term; all Rwandans should be umuturage and
they should be proud of that. In 1988, Ntamahungiro wrote:
To give a medal of honour to each and every peasant.To decorate some peasants as Model
Farmers. To give decorations at certain officials considered close to the peasantry. To
baptise a street, a place, a hotel, a day in the name of the peasants. To compose a song in
their honour. To organize popular parties in each commune or sector. There is no
shortage of ideas and we can count on the creativity of certain minds to supply
tailor-made expressions . . . We know, however, how much this part, the majority of the
population, suffers. The visits of the Minister of Internal Affairs and Communal Development and of the Minister of Justice have shown us some of these injustices. From her
side, the national press regularly provides evidence of the poverty in the rural areas and
in the cities. (Ntamahungiro 1988, 8)
In Habyarimana’s speeches, in MRND documents and in the writings of Rwandan authors and
scholars, ‘feudal’ always refers to the Tutsi monarchy who ruled Rwanda before the 1959–62
Revolution. In a 1987 Anniversary book commissioned by the President’s office, the 1959–62
Revolution is called a peasant revolution. In other works, it is called the Social Revolution or the
Hutu Revolution. This means that the term ‘peasant’ is used for ‘Hutu’ and the term ‘bourgeois’
or ‘feudal’ is used for Tutsi. In other words, in Habyarimana’s ideology the Tutsi were not
peasants; they were always considered the bourgeois or feudalists. This juxtaposition is clearly
demonstrated in the work of J.-P. Chrétien – a French historian – and Anna-Maria Brandstetter
– a German anthropologist who specializes in Central Africa:
The government presented itself as République égalitaire and continues to set its hopes on
the myth of the egalitarian, peasant society in spite of the growing social and economic
tensions. It looks upon itself as the inheritance of the ‘peasant revolution’ . . . The regime’s
10
The peasant ideology of the regime is explained in detail in Verwimp (2006). For a comparative approach to
the contribution of peasant ideology and ruralization to genocide, see Nairn (1998).
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 405
founding ideology spoke of the sociological majority (la pure démocratie du peuple majoritaire)
which had permanently overcome the ‘minority of the feudal Tutsi’ (minorité des féodaux
tutsi). The Hutu were equated with a democratic majority or ‘majority people’ (rubanda
nyamwinshi) and the Tutsi with an aristocratic and feudal minority . . . Rwanda was termed
‘the Land of the Hutu’ . . . and the opposition between Hutu peasantry and Tutsi feudalism
remained central to their discourse.11
Main Ideologues
The activities of three people, all closely connected with the top of the MRND and Habyarimana himself, deserve closer attention: Hassan Ngeze, Léon Mugesera and Ferdinand
Nahimana. The journal Kangura, which started publishing anti-Tutsi articles in May 1990, was
printed at a state-run printing press in Kigali, receiving subsidized credit or reduced prices
(African Rights 1995). In its June 1990 issue, four months before the start of the war, editor
Hassan Ngeze claimed that 70 per cent of Rwanda’s prominent businessmen were Tutsi. He
also wrote that these rich Tutsi collaborate with refugees outside the country Kangura (1990, 3).
The theme of Tutsi wealth and Tutsi control of the Rwandan economy would come up in
many editions of Kangura in subsequent months and years. In December 1990, two months
before the massacre of the Bagogwe (see below), the ‘Ten Commandments’ (part of an ‘Appel
à la conscience des Bahutu’) were published in Kangura. The text is a racist anti-Tutsi statement,
prescribing rules of behaviour for all Hutu in their interaction with Tutsi. It argues that all Hutu
who have Tutsi wives,Tutsi concubines and all Hutu who do business with Tutsi are traitors.The
Appel says that the Tutsi inside Rwanda are the accomplices of the RPF. These messages were
very effective in arousing fear of the Tutsi and brought home the explicitly stated assertion that
‘the enemy is among us’.
Ferdinand Nahimana was a university professor of history before he became director of
ORINFOR. He was already ‘active’ in 1973 on the Butare campus Committee du Salut Public,
which implemented the expulsion of Tutsi students and professors from the university. In 1988,
he published a book on Rwandan culture and development in which he explained and
glorified Habyarimana’s approach to development and to the peasantry. In his book, umuganda
is described as a virtuous practice, deeply ingrained in Rwandan culture and tradition.12 Under
his leadership, Radio Rwanda was openly racist. In February 1993, Nahimana advocated a
civilian defence force made up of young people. He stressed the usefulness of such a popular
force to safeguard the peace inside the country (Des Forges 1999, 110). In August 1993,
Nahimana became head of RTLM, the notorious hate radio.
Léon Mugesera was vice-president of the MRND in Gisenyi. In November 1992, he gave
a speech in Gaseke commune that would resonate nationally because of its racist content.13 In
the speech, he rhetorically asked whether the Hutu were waiting for the Inyenzi (cockroaches)
to come and kill them. They made a fatal mistake in 1959, he said, by allowing the Tutsi to
leave. If the judicial authorities do not act against RPF accomplices, he said, the population
must take matters into their own hands. He asked whether his audience knew that the Falasha
11
Brandstetter (1997) refers to Chrétien (1991, 1992) and Panabel (1995).
See Nahimana (1988). The author states that the book has been published with the support of the presidency
of the MRND and the Ministry of Education and Scientific Research. In 1990, an official MRND publication
even goes a step further when it deplores that the value of umuganda was lost through contact with the colonizer
and in particular because of the introduction of money, the generalization of education and salaried employment
(translated from MRND 1990, 10).
13
Speech by Léon Mugesera before the militants of the MRND, sous-préfecture de Kabaya, Gisenyi, 22
Novembre 1992 (mimeo, translation from French to English by the author).
12
© 2011 Blackwell Publishing Ltd
406 Philip Verwimp
(Ethiopia’s Jewish population) had returned to Israel and suggested that the Tutsi should be sent
to their homeland, Ethiopia, via the Nyabarongo river. He started and ended his speech with
a salute to President Habyarimana, who never disavowed the content or the speaker. He also
attacked Hutu of other parties who were negotiating with the RPF, and called them Inyenzi
talking with other Inyenzi. He blamed the Hutu of other parties for the loss of Byumba to the
RPF. Hutu should not allow themselves ‘to be invaded’. Mugesera ended his speech with a call
for unity: ‘we must all rise, act as one person’. Straus (2006, 197) interprets this speech as a call
for retaliation, self-defence and deterrence.
That Mugesera’s speech was understood as a ‘programme’ is confirmed by the words of the
burgomaster of Kibilira (see below), who said that that the programme announced by Mugesera
would be continued. Ngeze, Mugesera and Nahimana were seen by everyone as acting on
behalf of Rwanda’s leadership. They owed their jobs to their MRND mentors and were
die-hard ideologues of the cause. They incited hatred among the population with impunity,
which could only mean that they were protected from above.
ECONOMY
Despite the peasant-friendly rhetoric, the major policies enacted by the regime were or
remained peasant-friendly only when this did not hinder other objectives. Two examples
illustrate this. First, coffee policy. While the state-owned coffee marketing agency OCIR-Café
(Office des Cafés du Rwanda) gave a relatively high price to peasant producers in the first
half of the Second Republic, Habyarimana’s penal code contained penalties for ripping out
or neglecting coffee trees. When, from the second half of the 1980s onwards, the international price of coffee plummeted, the regime could no longer afford to buy the loyalty
from the peasant producer offering a high price and switched to coercion and repression to
maintain its power.14 Second, food policy. Habyarimana was very preoccupied with the foodpopulation equilibrium. Because he did not want to introduce family planning to reduce the
number of births, he could only try to increase food production. Production increased at the
rate of population growth as long as new land was taken into cultivation. Willame (1995,
135) writes that productivity, however, never increased. Thus, as soon as all land was taken into
cultivation, the production of food per capita started to decline. Notwithstanding a commitment to food production, Habyarimana did not hesitate to move people from their land,
undoubtedly because of the need for export earnings. In Gisovu commune, Kibuye province,
several hundred households were expropriated to make room for a tea plantation. Bart (1993,
456) writes that this met with open hostility from the population. Expropriation also
occurred in Gisenyi, where 450 households were removed from their land to make space for
the expansion of the tea plantations in Nyabihu in Karago commune.Von Braun et al. (1991,
114) write that about 300 hectares of land was expropriated for this purpose.15 They found
that the expropriation took place in 1977 and 1985. Displaced households, more than other
households, earned a larger part of their income by working for the tea factory (Von Braun
et al. 1991, 81). Not all of the displaced households remained in the commune. Bart (1993,
456) writes that a few dozen households were resettled in the paysannat of nearby Kinigi,
while others left for Bugesera or Mutara.
The mid-1980s marked a turning point in the state of Rwanda’s economy under Habyarimana.There was no more land to be taken into cultivation.The land size per household was 1.4
14
Verwimp (2003) describes in detail the political economy of coffee and power under Habyarimana.
On a national level, expropriation and conversion of land for tea estates took place on forested land, pasture
land and cultivated land (IFPRI 1991; World Bank 1991).
15
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 407
hectares at the end of the 1970s and only 0.8 hectares in 1991. Drawing on survey research in
1988 and 1993, André and Platteau (1998) have demonstrated the extent of land conflicts
between and within households. The international price of coffee plummeted in 1986 and in
1989. By 1990, the coffee farmer could buy, in real terms, only 1 kg of beans with 3 kg of coffee,
where this rate had been one-to-one ten years earlier. Rwanda’s sole mine (tin) was closed.
Growth of real GDP per capita declined from 1.7 per year at the end of the 1970s to -2.6 per
year ten years later (World Bank 1991; Berlage et al. 1993).The discontent of farmers was revealed
when they ripped out coffee trees and refused to show up for umuganda, the weekly communal
labour programme (OCIR-Café 1992;Tardiff-Douglin et al. 1993; Uvin 1998). Despite the fact
that food self-sufficiency was the central aim of Habyarimana’s agricultural policy, and indeed the
key declared objective of his entire reign, the southern prefectures of Gikongoro and parts of
Butare were hit by crop failure and famine in 1989. Instead of coming to the rescue of desperate
peasants, the regime forbade journalists to write about the crop failure, did not issue food import
licences until a year later and denied that starvation was happening in the south.16 The
mismanagement and eventual dismantling of state food agency OPROVIA, the marketing board
set up to protect farm-gate prices for beans and sorghum, which could have intervened to stop
the starvation, caused despair among the peasants (Pottier 1993).
Economic development was understood as maintaining the peaceful life of the peasant
population in the rural areas. The peasants’ only duty was to produce, as the leadership took
upon itself the burden of managing the affairs of the state (Prunier 1995).The MNRD was not
a political party; it was a movement for development. The Parliament was named the National
Council for Development. This approach to rural development was an integral part of the
ideological stand of the regime. This can be illustrated by the absence of rural-to-urban
migration. Rwanda was the least urbanized country in the world (95 per cent of the population
lived in rural areas) and the regime wanted this to remain that the case. This is time and again
repeated in Habyarimana’s discourse.While the rural character of Rwandan society was praised
by the authorities for making it possible for the capital not to have any slums, its consequences
were in fact dire: a massive concentration of the working population in the agricultural sector,
with access only to tiny plots of land, and without any hope of ever leaving agriculture for
either the current or the subsequent generations (Uvin 1998). In a document of the Ministry
of Planning, it is argued that ‘we should avoid that the unemployed rural masses come to the
city where they can cause social and political upheaval.’17 While the deliberate choice not to
urbanize Rwanda was underpinned by ideological and political motivations, it hindered economic development (World Bank 1991).
THE MASSACRES COMMITTED BETWEEN OCTOBER 1990 AND MID-1992
Mass Imprisonment Right after the Start of the War in October 1990
It is highly likely that Rwanda’s intelligence service informed Habyarimana of the upcoming
attack by the RPF (Adelman and Suhrke 1996, 20). Already in May 1989, at the Nyagatare
summit, the Ugandan president Museveni had warned Habyarimana of a potential invasion
(Nsengiyaremye 1995, 247). This meant that the regime could prepare itself for the attack. The
16
The vivid reality of crop failure and starvation was, at the time, revealed in the Catholic periodical Kinyamateka
in 1989 and 1990.This episode of crop failure and starvation, as well as the denial and the inaction of the regime,
have been analysed in detail in Verwimp (2002). The effects of the crop failure on child health are analysed in
Akresh et al. (2011).
17
Translated from the French text in Guichaoua (1988).
© 2011 Blackwell Publishing Ltd
408 Philip Verwimp
‘preparation’, however, was not a military one, as one would expect. Des Forges (1999, 49)
writes that the Rwandan commander at the border, aware of the pending invasion, demanded
reinforcements from headquarters. He got none, leading him and others to speculate that
Habyarimana wanted the invasion. Apart from soliciting French military support (and thus
drawing foreign powers into the conflict), few military preparations were made. Asked by the
author why the regime did not fight the rebels more forcefully (it is well known that the war
between the RPF and the FAR was a low-intensity war, with a limited number of battles and
a limited number of casualties), a person close to Habyarimana answered ‘I believe Habyarimana
counted on the population’.18 We encounter here the notion that ‘the population or the people’
will at some point come to the rescue of the nation, embodied in the figure of the president.
Only three days after the attack, the regime launched a faked attack on the capital Kigali,
which allowed it, under the veil of assuring security, to round up 8,000–10,000 people and put
them in prison.19 Many of them were Tutsi businessmen and intellectuals. They were held
without charge, in deplorable conditions, for several months; they were tortured and several
dozen died in prison.20 It was not the first time that the regime had rounded up many people
in Kigali. In the mid-1980s, it had launched a campaign against ‘loose women’ by taking young
girls from the street who were accused of being prostitutes and transporting them to a
re-education camp (Jefremovas 1991).
The Hima of Mutara (Savannah in North-east Byumba)
Until 1972 several groups of pastoralists, the Hima, lived as a nomadic people with their herds
of cattle in the savannah of Mutara, in the north of the prefecture of Byumba, in the communes
of Muvumba and Ngarama. In August and September 1973, 4,762 Burundian refugees were
installed in Mutara, near the church of Rukomo (Bart 1993, 397). Their settlement site was
known as the paysannat of the Barundi and it marked the beginning of the colonization of the
Mutara region. From 1974 onwards, OVAPAM (Office pour la Valorisation Pastorale et Agricole
du Mutara), a large integrated project, installed 11,850 families in an area of 37,000 hectares
(Bart 1993, 526). The pasture land was organized into ranches, where the cattle owners were
taught modern livestock techniques. In order to benefit from these services, the pastoralists had
to sign a contract in which they renounced their rights to other land and agreed to follow the
instructions given by OVAPAM staff for the treatment of the cattle.
On 8 October 1990, one week after the beginning of hostilities between the RPF and the
FAR, soldiers from the FAR murdered at least 65 Hima in Mutara.21 A journalist from De
Standaard (Belgium’s leading newspaper) visited Rukomo several days after the massacre. He
wrote that it was clear that the Hutu from the paysannat were implicated in the killings and in
the looting of the Hima ranches. The journalist added that his interviews with people in the
area contradicted official statements denying that there was a bloodbath (President Habyarimana) or blaming the Tutsi for the killings (the Rwandan Embassy in Washington, DC). He also
wrote that people in Rukomo believed that the killing was planned ahead, and they considered
18
Interview, Kigali, November 2000.
Straus (2006) doubts that the faked attack was intentional and argues that it may have been caused by panicky
soldier firing (p. 192).
20
De Standaard, 15 October 1990; see also Reyntjens (1994, 95) and Desforges (1999, 49).
21
ADL (1992, 83–5). The report mentions many other names, but without exact dates of death, which is the
reason why I have written at least 65. In Hope for Rwanda, Sibomana (1999) put the figure at several hundred, a
number also used by FIDH (1993, 62) and Des Forges (1999, 50).
19
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 409
it as punishment for the Hima because they were believed to have aided the RPF.22 Prunier
(1995, 138) writes that these killings were preceded by a radio message from the Minister of
Defence, demanding that the population ‘trace and arrest those who infiltrate’. An officer of the
FAR, interviewed by the International Federation of Human Rights Organisations (FIDH), said
that several FAR companies were given the order to clean the zone between Nyagatare and
Kagitumba of all its inhabitants (FIDH 1993, 62).
The Tutsi of Kibilira Commune (Gisenyi)
In the second act of mass murder in mid-October 1990, 348 civilians were killed in 48 hours in
Kibilira commune in Gisenyi province (FIDH 1993, 20).The report is very clear on the role of
the communal authorities.They incited the population under the fabricated story that Tutsi had
come to exterminate Hutu. The burgomaster, who was taken to prison (and released several
weeks later) for his role in the mass murder, declared that people should ‘continue working’.
Independent witness accounts confirm the role played behind the scenes by Léon Mugesera who,
two years later, would deliver racist speeches in Kibilira and the neighbouring Gaseke commune.
One of the local government agents (conseiller) directing the slaughter declared to the investigators
that he had followed the attackers to guarantee their security.The same Tutsi families in the same
commune would fall under attack again in March 1992, at the same time as the massacre in
Bugesera (see below), and again in December 1992. On 10 January 1993, the burgomaster of
Kibilira said that the programme announced by Mugesera had not changed and would resume
when the international investigators (who were in Rwanda in January 1993) had left.
The massacre in Kibilira happened one week after 8,000–10,000 people were taken from the
streets and imprisoned in Kigali, in the first week of October 1990.Thus, already from the very
beginning of the civil war, in the rounding up of many people in the capital and in the two
massacres committed in October 1990, we encounter several ingredients that were to characterize
subsequent massacres. First, attacks were fabricated and stories were spread to allow the regime
to rally support, undertake an operation and incite the population to kill Tutsi civilians; second,
the authorities (national or local) took the lead; third, these same authorities lied about the nature
of the operation and denied that one ethnic group was targeted; fourth, the operation was
legitimated under the veil of assuring security; fifth, the metaphor of ‘work’ was used to describe
the killing; and, sixth, national-level figures or ideologues monitored the local campaign.
Only at a later stage, towards the end of 1992 and in 1993, would so-called Hutu moderates
also be killed in targeted attacks. This sequence is important because some scholars argue that
after 6 April 1994, Hutu moderates were the first to be killed. The examination of what took
place prior to 1994, however, shows that this provides an incomplete and inaccurate picture of
the sequence of events.
The Bagogwe of the North-west (Gisenyi and Ruhengeri)
Between 25 January 1991 and 4 February 1991 (three years before the genocide), a massacre
was carried out against a group of Tutsi known as Bagogwe.They used to be – and for the most
part still were in 1991 – pastoralists. The Bagogwe preferred to live in the high mountainous
regions, with good pastures for their cattle. Only recently, with the reduction of pasture land,
had they begun to cultivate. At least 300 people (and a maximum of 1,000; FIDH 1993, 37)
were killed in a series of brutal attacks in several sectors of the north-west of the country, in
22
De Standaard, 13–14 October 1990, p. 2.
© 2011 Blackwell Publishing Ltd
410 Philip Verwimp
the prefectures of Gisenyi and Ruhengeri. According to the 1993 report, President Habyarimana himself presided over the meeting that organized the massacre of the Bagogwe:
The journalist Janvier Africa worked as an agent for the Central Information Service until
the beginning of the war, after which he worked directly for the Presidency. He confirms
that he assisted in reunions held by a group known as ‘ Death Squads’ (Escadrons de la
Mort). He recalls a reunion at 2.00 am in January 1991 before the attack on Ruhengeri
by the FPR. Participating in this reunion were Joseph Nzirorera (then Minister of Mines
and Handycraft), Charles Nzabagerageza (then préfet of Ruhengeri), Côme Bizimungu
(then préfet of Gisenyi) and Casimir Bizimungu (then Minister of Foreign Affairs). After
the liberation of the city, they decided to kill the Bagogwe. Colonel Sagatwe, Protais
Zigiranyirazo (brother-in-law of the president), member of parliament (député) Rucagu
and préfet Nzabagerageza all agreed on that point. Préfet Nzabagerageza should instruct
burgomasters to find trustworthy people to do the job. Janvier Africa confirms that it was
a big operation that cost 15 million Rwandan francs.The role of Janvier was to verify the
results of the operation, to make sure that those who had to be killed really were dead.
He showed credible evidence of his participation in the operation.
The reunion that prepared the massacre of the Bagogwe was presided by Juvenal
Habyrimana himself, his wife was also present, as well as Colonel Sagatwa and his wife
and a traditional truth-teller (sorcier) invited by Sagatwa. Minister Joseph Nzirorera was
charged with the delivery of the money to préfet Nzabagerageza.
It was Colonel Elie Sagatwa who proposed the massacre of the Bagogwe and President
Habyarimana agreed by nicking his head. Nzirorera, Nzabagerageza and Côme Bizimungu had to look for trustworthy Burgomasters. Once the operation started, one had to
make sure that the police participated in order to get the job done. (FIDH 1993, 38;
author’s translation from French text)
When reading about the preparation of the massacre, it is clear that this massacre was not a
spontaneous outburst by an anxious population. It was planned and organized by the national
leadership.The fear of the RPF was twisted and manipulated by the leadership into an immediate
threat to Hutu livelihoods, thereby inducing the Hutu population ‘to act first’. A fake assault –
fabricated to legitimize the campaign – worked so well that the immediate reaction of the Hutu
population was to flee.The burgomaster had to persuade them to stay and attack their Bagogwe
neighbours (Des Forges 1999, 88). Since the massacre of the Bagogwe was executed right after
an attack by the RPF on the centre of Ruhengeri, it seems easy to infer that the massacre was
an act of retribution (or revenge) by the Habyarimana regime. However, the advocates of the
revenge interpretation fail to explain why the revenge took the form it did; that is, the massacre
of unarmed civilians. Revenge could have taken several other forms, such as killing Tutsi who
were still in prison after the October 1990 raids in Kigali or launching an offensive against the
RPF.Throughout the civil war, the regime spent a lot of energy attacking and killing the unarmed
Tutsi civilian population inside Rwanda.This is what needs to be explained and ‘revenge’ is far
from accomplishing that.The civil war indeed accounts for the timing of the massacre, but it does
not explain why these massacres took the form of ethnic cleansing.
The Tutsi of Bugesera (South-central Rwanda)
In March 1992, authorities organized the killing of several hundred Tutsi in Bugesera, a region
located to the south of Kigali where Hutu (from the north-west) and Tutsi (from the
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 411
south-west), both from densely populated areas in Rwanda, had recently migrated and settled.
Figure 2 shows how Hutu and Tutsi from the north, but in particular from the south of
Rwanda, migrated to the new lands in Bugesera (the communes of Ngenda, Kanzenze and
Gashora) that were to be colonized.
Hassan Ngeze, editor of Kangura, visited the area several times prior to the massacre and
spread tracts and rumours about the danger of the Inyenzi (Des Forges 1999, 89). On 3 March,
Radio Rwanda issued a warning that Tutsi were going to kill Hutu; in particular, Hutu leaders
in Bugesera. At that time, Ferdinand Nahimana was director of the Rwandan Office for
Information (ORINFOR), where he supervised Radio Rwanda. The burgomaster of Kanzenze, Mr Fidèle Rwambuka, who played a leading role in the massacre, was a member of the
Central Committee of the MRND. Rwambuka, who denied knowing about the massacre
when interviewed by the FIDH, could count on the support of interahamwe (at that time, the
name for the youth militia of the MRND) despatched from Kigali and on soldiers from the
nearby Gako camp. In Nyamata in Bugesera, one can visit the grave of Sister Locatelli, an Italian
nun who was living there at the time of the massacre. She warned embassies in Kigali that the
massacre was taking place and was subsequently killed by the perpetrators.
An experienced observer of Rwanda’s history, David Newbury, described the period as
follows: ‘With the pretence of looking for internal enemies, from late 1990 and early 1991,
there were small-scale killings (of several hundred people) and wider roundups of “suspects”
within Rwanda. The military leaders learned two principal lessons from this exercise: that such
tactics were feasible, and that they generated no meaningful response by outside powers’
(Newbury 1998, 79).
THE POLITICAL GEOGRAPHY OF THE MASSACRES
A clear pattern characterizes the massacres in Kibilira, of the Bagogwe, and in Bugesera, as
follows. Fabricated stories are spread, stating that Tutsi have killed or plan to kill Hutu;
ideologues are present at massacre sites to give speeches or animate meetings; ‘trustworthy’
burgomasters are enlisted to call meetings with the conseillers; young people and interahamwe
are despatched to hunt, pillage and kill. Each time, FIDH and ADL establish a personal and
organizational link with the national leadership in Kigali, making these massacres all but
spontaneous outbursts of violence. Most of these massacres took place in the north-west of
the country, where the MRND was strongly supported by the local administrators and the
population.23 Habyarimana’s home region (Gisenyi and Ruhengeri) had received by far the
largest amounts of government subsidies and had benefited from the greatest number of
government jobs. The other area where a massacre took place was Bugesera, in the rural part
of the province of Kigali, a region that had only recently been populated by Hutu from Gisenyi
and Ruhengeri, as well as by Tutsi from Gikongoro and Butare.
The strong support for the MRND is not the only element that distinguishes these
provinces and communes. Gisenyi and Ruhengeri are by far the most densely populated
provinces in Rwanda. In 1991, accounting for the forested areas in both provinces, Gisenyi
accounted for 735,000 people on 1,350 km2 of cultivable land, which is 560 persons per km2,
and Ruhengeri 532 persons per km2.This is almost twice the average for the other provinces.24
23
Out of 17 incidents of serious violence in the 1990–93 period, 14 took place in Gisenyi or Ruhengeri (Des
Forges 1999, 87).
24
Only urban centres had an even higher density, but there people do not live off the land. Gisenyi is also the
location of the fieldwork undertaken by C. André in 1988 and 1993. She found extreme pressure on land, which
even deteriorated in just five years. This resulted in many conflicts over land and a large number of landless or
© 2011 Blackwell Publishing Ltd
412 Philip Verwimp
The average size of a farm in Gisenyi (0.45 hectare) was by far the smallest compared to the
other provinces. The communes in Gisenyi where the violence against Tutsi was orchestrated,
Kibilira, and in later instances of violence also the commune of Mutura, have the highest
percentages of Tutsi (8.6 and 9.7 per cent of the population, respectively) in the province. For
the province of Ruhengeri, the communes of Kinigi (3.7 per cent) and Mukingo (2.1 per
cent), where Bagogwe were killed, had the highest percentage of Tutsi in Ruhengeri. Kanzenze
commune, in recently settled Bugesera, had the highest percentage of Tutsi in the entire
province of Rural Kigali (31 per cent).25 Recall that the Hima of the savannah in northern
Byumba were a pastoral people, who had recently been settled into ranches as part of a large
land resettlement programme. Thus, these first massacres occurred in places combining the
following features:
• strong MRND support among local authorities and/or the population;
• the most densely populated areas in rural Rwanda or recently (re)settled areas; and
• communes chosen that had highest the percentage of Tutsi in the province.
In other parts of Rwanda, a smaller number of Tutsi were killed (compared to the four
massacres described above) in the 1990–92 period. In those places, we find evidence of the
same logic as in the massacre sites above:
• The communes of Rwamatamu and Gishyita, in Kibuye province: a higher than average
population density and a very high percentage of Tutsi.
• The region of Nasho in the Rusumo commune, Kibungo province: a pastoralist population
•
•
settled after the 1982 expulsion from Uganda. Pastoralists killed by FAR soldiers and
members of the local paysannat.
The Rwanbuka sector in the Murambi commune, Byumba province: an MRND stronghold, with the burgomaster originating from the sector where the killings occurred.
The Mugina commune, Gitarama province: the killing of a Tutsi specifically to occupy his
pastoral land.26
What is evident is that immediately after the beginning of the October 1990 civil war, Tutsi
were targeted and killed in local massacres. More specifically, the places where these massacres of the Hima, Bagogwe and Tutsi were committed in the 1990–92 period were located
on the land frontier, in recently settled areas or in paysannat settlement schemes. Several of
the places where the early massacres occurred, such as in Mutara (northern Byumba), Kinigi,
Mukingo, Bugesera and Rusumo (see Figure 3), were places where paysannats had been
established. In fact, ten out of the 19 communes where massacres occurred in the period
1990–92 were communes with paysannat settlement schemes (see Table 1 and Figure 4). As
Rwanda had 17 communes with paysannats, which is 12 per cent of the total number of 145
communes, this means that 59 per cent (ten out of 17) of the communes having such a
settlement scheme were hit by a massacre. This figure needs to be compared with the
probability of a massacre in communes without paysannats, which was 7 per cent (nine out
quasi-landless peasants, whose farm size was too small to make a living, to feed the family and to offer land to sons
who wanted to marry.
25
The source of these percentages is the 1983 count of the population by ethnicity in the administrative records
kept at the commune level.
26
See ADL (1992). This report uses the word ‘genocide’ on several occasions to describe the massacres. See also
FIDH (1993, 52–5).
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 413
Figure 3 The paysannat of Rusumo, province of Kibungo
1 km
Akage
ra
towards
Kigali
Road, track
Rugo
Cultivated space
Space being cleared
for cultivation
Marsh
towards
Tanzania
Source: Bart (1993).
© 2011 Blackwell Publishing Ltd
414 Philip Verwimp
Table 1. The number of communes (N = 145), communes with a paysannat, and communes with a
paysannat and a massacre, 1990–92
Communes with
a paysannat
Communes with a
massacre, 1990–93
Communes with a
paysannat and a massacre
17 out of 145a
12%
19 out of 145b
13%
10 out of 17c
59%
a
Bart (1993, 394) with at least 1,000 households.
Des Forges (1999, 87), Reyntjens (1994, 186) and ADL (1992).
c
These ten communes are: Muvumba in northern Byumba; Mukingo, Nkuli and Kinigi in
Ruhengeri; Mutura and Karago in Gisenyi; Kazenze, Gashora and Ngenda in Rural Kigali; and
Rusumo in Kibungo.
Pearson chi-square test (1) = 35.35, with p = 0.000.
b
Figure 4 Paysannat and massacre sites in Rwanda, 1990–92. A circle represents a paysannat
Source: Bart (1993). A cross represents a massacre; sources as for Table 1.
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 415
of 128), an enormous difference.27 In Gashora – a commune neighbouring Kanzenze in
Bugesera – where part of the March 1992 killings occurred, a model village was established
for model farmers, with rectangular houses built in a grid of straight roads.
In the former paysannat visited during the fieldwork in Gitarama in 2004 mentioned above,
the agronomist, the agricultural surveyor and the agricultural monitors became the leaders of
the killing operations in the 1994 genocide. The agents of the developmental state become the
perpetrators of genocide.
In Kanzenze (Bugesera), the commune most touched by the March 1992 massacre, two
thirds of the population lived in a paysannat (Bart 1993, 382). The early massacres inscribe
themselves in a logic of land colonization, resettlement, depredation and deprivation of cattle
and land in areas where the land constraint was biting most severely and where peasant society
was being remodelled in a rational, geometric way.28 These early massacres can thus be
described and understood as acts of spatial and social engineering through ethnic cleansing: the
removal of pastoralist groups from the land in order to occupy the land for food cultivation,
paysannat settlement schemes and export crop production. In times of civil war, the Tutsi need
not be resettled; there is no space for them anyway, so they can be killed.
Just as there was no longer any space for pastoralism after the 1959 revolution and no space
for Tutsi refugees in the 1970s and 1980s, there was now no longer any space for Tutsi in
Rwanda. Delicate operations such as massacres could, at that time, only be executed in MRND
strongholds. The massacres occurred in strongholds with very high population pressures, and
high percentages of Tutsi compared to the provincial level, and in areas with previous experience of land colonization and resettlement such as the paysannats. They were executed as
umuganda, the obligatory communal labour.The policy of communal labour was introduced by
Habyarimana in 1975 to re-establish the value of manual labour. It gave local authorities a lot
of sway over the peasant population, which they used to mobilize people for the killings.
CONCLUSION
This paper has put forth a third, alternative interpretation of the massacres that were perpetrated
in Rwanda prior to the 1994 genocide. This interpretation, which is based on the PeasantState’s ideology of excluding the pastoralist groups from the realm of the state, remodelling the
agrarian space and colonizing new land, challenges, or at least complements, the two prevailing
interpretations in the literature.
The Bahima and the Bagogwe had not sought to take on positions of power in the state;
indeed, they had sought to stay away from the influence of state power. Prunier (1995, 169)
writes that, in the context of Rwanda, the Bagogwe were poorer than average. They lived off
their cattle on the little pastoral land that was left. However, the state would not leave them
alone. In the commune of Mukingo, in the north-west of the country, a paysannat was
established on their land, whereby contracts were signed with farmers with a view to the
growing of pyrethrum for export. In the north-east, ranches were created for their cattle as part
of a large-scale rural development programme. Whatever the specific form took in a particular
case, the point is that no one could escape the Second Republic’s drive to register, handle,
27
The comparison becomes 45 per cent (ten out of 22) compared to 7.3 per cent (nine out of 123) when
including the very small paysannats located in the provincial capitals. The chi-square statistic is 35.35, with one
degree of freedom and a p-value of 0, meaning that the result cannot be ascribed to chance.
28
Just how tough the effects of the land constraints were can best be understood in a paper by C. André and
J.-Ph. Platteau (1998). The fieldwork for their paper was undertaken in 1988 and 1993 on a hill in Gisenyi
province. Verwimp (2005) presents an economic profile of the perpetrators.
© 2011 Blackwell Publishing Ltd
416 Philip Verwimp
monitor and develop its people. The 1990–92 massacres took place in communes where there
was already a substantial experience of spatial and social engineering on the part of the Second
Republic, and where agrarian space had been significantly remodelled from above.Then, in the
context of the civil war, spatial and social engineering went a step further, meaning that the
pastoralists were killed rather than resettled. The fate of Rwanda’s Bahima and Bagogwe
pastoralists illustrates the nature of the developmental state: ‘seeing like a state’, the regime
decided that these pastoralists should be removed from the body politic. The cover of the civil
war was thus used to rid specific areas of their pastoral inhabitants.The regime denied that any
massacres had taken place, and instead described the events as either spontaneous outbursts of
violence (when far from the battle front) or as war operations (when close to the front).
The Habyarimana regime had used up all the available land in Rwanda. The regime had
reached this point by way of an extensive land strategy – turning land that was used for pasture
into agricultural land and colonizing new (often marginal) land, mainly through the paysannat
settlement scheme. The primacy of agriculture over pastoralism was a key outcome of the
Hutu Revolution, as shown by the abolition of Ibikingi rights and the subsequent cultivation
of pastoral lands.The Rwandan state intervened strongly in land management, not least through
the establishment of paysannats and the expropriation of households with a view to the creation
of tea plantations.This policy was met with resentment, given that farmers on the land frontier
were forced to move and ended up with smaller parcels. In 1986, the Central Committee of the
MRND refused the return of Tutsi refugees from Uganda, arguing that there was no space for
them. In 1989 and 1990, Habyarimana maintained that many of the country’s problems, such
as famine, were the result of population growth.
The Habyarimana regime adopted a policy of agricultural extensification, as opposed to
intensification. This meant turning all available land (such as pastures, marches and forests) into
cultivable land. In Rwanda, this policy came up against its limits in the late 1980s, as all the land
had by then been taken over. It is therefore important to note that, by 1991, most Bagogwe were
still pastoralists.They preferred to live in the high mountainous regions, where there were good
pastures for their cattle. Only more recently, with the reduction in pastoral land, had they begun
to cultivate. Pastoralists such as the Hima and the Bagogwe live off cattle and the trade in cattle
products.They do not cultivate and are therefore considered a non-agricultural group. Pastoralism
as a way of life did not fit into the agrarian order of the Second Republic, which was based on
hard-working smallholder peasants.The regime used the opportunity provided by the civil war
in order to claim the last remaining parcels of land by removing the last remnants of pastoralism
in Rwanda. This was perfectly in line with the prevailing ideology of the Second Republic; in
other words, that Rwanda was and would always remain an agrarian nation of hard-working
peasants.The point was not that the Hima, the Bagogwe and theTutsi owned cattle (for many rich
Hutu and the dignitaries of the Habyarimana regime did so, too), but that their pastoralist
livelihood did not fit in the Peasant-State. Peasant cultivation and rural life can be romanticized,
but when this ideology is combined with racism towards ethnic groups regarded as non-peasant,
such as the Tutsi, this ideology can take on a vicious character.29 Thus, in view of all of the above,
these massacres can be adequately described as a case of ethnic cleansing.
The international commission that wrote the FIDH report of March 1993 discussed the
applicability of the term ‘genocide’ in the case of the massacres that it described in detail.Then,
in a report dated August 1993, the United Nations Special Rapporteur on Summary, Arbitrary
29
A strong example of the ideology in the mind of the organizers of genocide is found in an essay written by
Colonel T. Bagosora after the genocide (Yaoundé, October 1995). He writes that the civil war was an ethnic war
of Hutu against Tutsi, and that the Tutsi are a nilothic people of immigrants, without a country of their own.They
have tried to impose their supremacy on the rightful original inhabitants.
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 417
and Extrajudicial Executions concluded that the killings did indeed constitute an instance of
genocide according to the terms of the 1948 Convention on Genocide.
This paper has argued that the intention to kill a group of people for who they are (the core
element of the definition of genocide according to the 1948 Convention) was clearly present
from the beginning of the civil war: the Bahima, the Bagogwe and the Tutsi were targeted because
of their pastoralist character and because they were regarded as belonging to a different (NiloHamitic) race.The leaders of the Second Republic espoused a racial ideology and the acts of mass
murder that they perpetrated cannot be properly understood outside this racist paradigm.
President Habyarimana denied that any massacres had taken place in the case of both the
Hima (in October 1990) and the Bagogwe (in August 1991), just as a year earlier he had denied
that there had been a famine in Gikongoro.30 This denial of harm, suffering or killing is a part
of the classic repertoire of the perpetrators of genocide. Genocide is the result of a gradual
policy involving identification, hate propaganda, the militarization of society, resource allocation
and so on.The 1993 FIDH report revealed the existence of a high-level committee behind the
mass murders. This group met on several occasions, enabling its members to address a number
of issues and organize the subsequent events. The meeting at which the massacre of the
Bagogwe was decided also discussed the means that were to be used in the operation (FIDH
1993, 38). These means consisted of trustworthy burgomasters, 15 million RWF and the help
of police officers. By the time that the leaders of the regime took the decision to go ahead and
execute the mass murders, they already had a pretty good idea of how to do it.
This paper has sought to highlight the social and geographical features of the massacres
that were perpetrated in Rwanda in the early 1990s. These features, along with the available
evidence on the motives and organization behind the violence, are clearly indicative of an
agrarian logic underlying the massacres, which cannot be satisfactorily explained as a defensive
reaction to imminent invasion or as preparation for a wider genocidal project. Rather, to a
significant extent these massacres were rooted in the crisis of an ideological programme and its
associated policies. The massacres took place in locations chosen for their particular characteristics, the strong level of support enjoyed by the MRND, their very high population density,
the fact that there was a relatively high percentage of Tutsi, and their recent history of land
colonization or of remodelling of the agrarian space into paysannat settlement schemes. Under
the cover of the civil war, it was here that the regime unveiled its darker side: that of a
Peasant-State unleashing its full violent potential against people considered to be non-peasants.
REFERENCES
Adelman, H. and A. Suhrke, 1996. Joint Evaluation of Emergency Assistance to Rwanda. Copenhangen: Danida.
ADL (Association Rwandaise pour la Défense des Droits de la Personne et des Libertés Publiques), 1992. Rapport sur
les Droits de l’Homme au Rwanda. Kigali: ADL.
African Rights, 1995. Rwanda: Death, Despair and Defiance, 2nd edn. London: African Rights.
Akresh, R., P. Verwimp and T. Bundervoet, 2011. ‘Crop Failure, Civil War and Child Stunting’. Economic Development
and Cultural Change, forthcoming.
André, C. and J.P. Platteau, 1998. ‘Land Relations under Unbearable Stress: Rwanda Caught in the Malthusian Trap’.
Journal of Economic Behaviour and Organization, 34 (1): 1–55.
Bart, F., 1993. Montagnes d’Afrique, Terres Paysannes, le Cas du Rwanda. Bordeaux: Presse Universitaire de Bordeaux.
Berlage, L., H. Eyssen, M. Goedhuys, L. Sleuwagen and D.V.d. Bulcke, 1993. Rwanda: Disequilibrium, Reform and the
Manufacturing Sector. Country Background Paper. Washington, DC: The World Bank.
Bézy, F., 1990. Rwanda: Bilan Socio-économique d’un Régime. 1962–1989. Louvain: Institut d’Etude des Pays en
Développment – Université de Louvain-la-Neuve.
30
For the denial of the massacre of the Hima, see De Standaard, 13–14 October 1990, p. 2. For the denial of the
massacre of the Bagogwe, see Des Forges (1999, 90–1).
© 2011 Blackwell Publishing Ltd
418 Philip Verwimp
Brandstetter, A.M., 1997. ‘Ethnic or Socio-economic Conflict? Political Interpretations of the Rwandan Crisis’.
International Journal on Minority and Group Rights, 4 (3–4): 427–49.
Brandstetter, A.M., 2001. ‘Die Rhetorik von Reinheit, Gewalt und Gemeinschaft: Bürgerkrieg und Genozid in
Rwanda’. Sociologus, 51: 48–84.
Chrétien, J.-P., 1991. ‘ “Presse Libre” et Propagande Raciste au Rwanda: Kangura et “les 10 Commandements du
Hutu” ’. Politique Africaine, 42 ( June): 109–20.
Chrétien, J.-P., 1992. ‘La Crise Politique Rwandaise’. Genève-Afrique, 30 (2): 121–40.
Des Forges, A., 1999. Leave None to Tell the Story. New York: Human Rights Watch.
FIDH (International Federation of Human Rights Organisations), 1993. Rapport de la Commission d’Enquête sur les
Violations des Droits de l’Homme au Rwanda depuis le 1er Octobre 1990. Paris.
Guichaoua, A. 1988. Conservation sociale et emergence contrainte des centres urbains dans les Etats-paysans des
Hautes Terres Centrales. Espace, Populations, Societes, 2, 245–60, Villeneuve d’Ascq, France.
Guichaoua, A., 1991. ‘Les Travaux Communautaires en Afrique Centrale’. Revue Tiers Monde, XXXII (127): 551–
73.
Guichaoua, A., 1997. Les Antécédents Politiques de la Crise Rwandaise de 1994. Rapport d’expertise rédigé a la demande
du Tribunal Penal International des Nations Unies sur le Rwanda. Arusha.
Habyarimana, J., 1974. ‘Discours, Messages et Entretiens’. Kigali: ORINFOR.
Jefremovas, V., 1991.‘Loose Women,Virtuous Wives and Timid Virgins: Gender and Control of Resources in Rwanda’.
Canadian Journal of African Studies, 25 (3): 378–95.
Kangura, 1990. Rwandan monthly magazine, no. 3, June. Kigali.
Kiernan, B., 2007. Blood and Soil. New Haven, CT: Yale University Press.
Lemarchand, R., 2002. ‘Disconnecting the Threads: Rwanda and the Holocaust Reconsidered’. Journal of Genocide
Studies, 4 (4): 499–518.
Little, P.D. and M. Horowitz, 1987. ‘Agricultural Policy and Practice in Rwanda’. Human Organisation, 46 (3): 254–9.
Little, P.D. and M. Horowitz, 1988. ‘Agricultural Policy and Practice in Rwanda’. Human Organisation, 47 (3): 271–3.
Mann, M., 2005. The Dark Side of Democracy: Explaining Ethnic Cleansing. Cambridge, UK: Cambridge University Press.
Mazower, M., 1998. The Dark Continent: Europe’s Twentieth Century. New York: Knopf.
McDoom, O., 2009. The Micro-politics of Mass Violence:Authority, Security, and Opportunity in Rwanda’s Genocide. Doctoral
dissertation, London School of Economics.
MRND (Mouvement Révolutionaire Nationale pour le Dévéloppement), 1990. Umuganda dans le Developpement
National. Kigali: MRND, p. 10.
Nahimana, F., 1988. Conscience Chez-nous, Confiance en Nous: Notre Culture est la Base de Notre Developpement
Harmonieux. Ruhengeri: Presse National du Rwanda.
Nairn, T., 1998. ‘Reflections on Nationalist Disasters’. New Left Review, I (203): 145–52.
Nathan, L., 2005. ‘The Frightful Inadequacy of Most of the Statistics: A Critique of Collier and Hoeffler on Causes
of Civil War’. Discussion Paper 11. Crisis States Research Centre, London School of Economics.
Newbury, C., 1992. ‘Rwanda: Recent Debates on Governance and Rural Development’. In Governance and Politics in
Africa, eds M. Bratton and G. Hyden, 193–219. Boulder, CO and London: Lynne Rienner.
Newbury, D., 1998. ‘Understanding Genocide’. African Studies Review, 41 (1): 73–97.
Nsengiyaremye, D., 1995. ‘La Transition Démocratique au Rwanda (1989–1993)’. In Les Crises Politiques au Burundi et
au Rwanda (1993–1994), ed. A. Guichaoua, 239–63. Université de Lille/Paris: Karthala.
Ntamahungiro, J., 1988. ‘Eloge du Paysan Rwandais’. Dialogue, 130: 5–16.
OCIR-Café, 1992. Proces-verbal de la Reunion du Conseil d’Administration de l’Office des Cafés, Mars 25. Kigali.
Olson, J.M., 1990. The Impact of Changing Socio-economic Factors on Migration Patterns in Rwanda. PhD dissertation, Ann
Arbor, Michigan State University.
Panabel, J.-P., 1995. ‘Bilan de la Deuxième République Rwandaise: Du Modèle de Développement à la Violence
générale’. Politique Africaine, 57 (mars): 112–23.
Pinchotti, S. and P. Verwimp, 2007.‘Social Capital and the Rwandan Genocide: A Micro-level Analysis’. Working Paper
30. Brighton: Households in Conflict Network.
Platteau, J.P., 2000. Institutions, Social Norms and Economic Development. Amsterdam: Harwood.
Pottier, J., 1993. ‘Taking Stock, Food Marketing Reform in Rwanda 1982–1989’. African Affairs, 92 (366): 5–30.
Prunier, G., 1995. The Rwanda Crises: History of a Genocide. New York: Columbia University Press.
Raleigh, C., A. Link, H. Hegre and J. Karlsen, 2010. ‘Introducing Acled: An Armed Conflict Location and Event
Dataset: Special Data Feature’. Journal of Peace Research, 47: 651–60.
Reyntjens, F., 1994. L’Afrique des Grands Lacs en Crise 1988–1994. Paris: Karthala.
Salih, M., T. Dietz and A.G.M. Ahmed, 2001. African Pastoralism, Conflict, Institutions and Government. London: Pluto
Press.
© 2011 Blackwell Publishing Ltd
The 1990–92 Massacres in Rwanda 419
Scott, J., 1998. Seeing Like a State, How Certain Schemes to Improve the Human Condition Have Failed. New Haven, CT:
Yale University Press.
Sibomana, A., 1999. Hope for Rwanda: Conversations with Laure Guilbert and Hervé Deguine. London: Pluto Press.
Straus, S., 2006. The Order of Genocide: Race, Power, and War in Rwanda. Ithaca, NY: Cornell University Press.
Tardiff-Douglin, D., J.L. Ngirumwami, J. Shaffer, A. Murekezi and T. Kampayana, 1993. Aperçu sur la Politique Cafeicole
á Rwanda. Kigali.
United Nations, 1993. Special Rapporteur’s Report on Rwanda, August. New York: United Nations.
US Department of State, 1993. Country Reports on Human Rights Practices, Report submitted to the Committee on Foreign
Affairs of the House and the Senate, February. Washington, DC.
Uvin, P., 1998. Aiding Violence: The Development Enterprise in Rwanda. West Hartford, CT: Kumarian Press.
Verwimp, P., 2002. ‘The Ruriganiza Famine in Southern Rwanda: A Prelude to Genocide?’ Discussion Paper 02.07.
Centre for Economic Studies, Leuven.
Verwimp, P., 2003. ‘The Political Economy of Coffee, Dictatorship and Genocide’. European Journal of Political
Economy, 19: 161–81.
Verwimp, P., 2005. ‘An Economic Profile of Perpetrators of Genocide’. Journal of Development Economics, 77: 297–323.
Verwimp, P., 2006. ‘The One Who Refuses To Work Is Harmful To Society’. First published in 1999 as a Working
Paper of the Genocide Studies Program, Yale University and in 2000 in the Journal of Genocide Research, 2 (3):
325–61. The 2006 version is online at www.hicn.org
Von Braun, J., H.d. Haen and J. Blanken, 1991. ‘Commercialisation of Agriculture under Population Pressure, Effects
on Production, Consumption and Nutrition in Rwanda’. Research Report 85. Washington, DC: International Food
Policy Research Institute.
Willame, J.C., 1995. Au Sources de l’Hécatombe Rwandaise. Paris: Institut Africain, Editions l’Harmattan.
World Bank, 1991. Rwanda: Agricultural Strategy Review. Washington, DC: The World Bank.
© 2011 Blackwell Publishing Ltd
