René Lemarchand

The threats posed to the state in both instances are inseparable from the introduction of the vote, and more generally from the ethos of democracy. The collapse of their state systems can best be seen as the ineluctable outcome of a head-on collision between the "premise of inequality" inherent in their traditional value orientation and the egalitarian message of liberal democracy. In vertically structured, minority-dominated societies, the verdict of the polls is never neutral, any more than the state systems to which they give birth. Rather than a "society-wide epiphany," to use Thomas Carothers' phrase (6), the result is violent conflict.

It is one thing to admit the potential for violence inherent in the electoral process, and quite another to gauge the scale of ensuing conflict. What needs to be underscored is that both states have experienced violence on a genocidal scale, and in both cases violence has generated massive outflows of refugees to neighboring territories. Theories of state collapse make little or no mention of the implications of genocide, both as an empirical fact and a phenomenon that profoundly alters the perceptions that one group has of the other. A notable exception is Alex De Waal's (7) lucid commentary on "the genocidal state": "Rwanda," he wrote,

is more than another collapsing state. The interim government of Rwanda is fighting for the right -- as it sees it -- to free itself from the moral claims of the rest of the world. This requires not just the eradication of the Tutsi minority but the annihilation of the human-rights and democracy movement in Rwanda, and all the values it stands for. In this furnace extremist politicians are reforging the identity of the Hutu people. It is frightening to watch.

In a society exonerated of moral constraints, and where the capture of power implies domination of one group by another, killing becomes a moral duty. The preservation of ethnic hegemony is perceived as a condition of physical survival, and the elimination of rival claimants the only means by which survival can be assured. Conversely, in such circumstances, the excluded community feels free to retaliate in kind. "An eye for an eye" becomes a license to kill. The result is endless bloodshed. In this hellish universe of mutually inflicted mass murder no one can claim innocence, nor is there any room for reconciliation and compromise. As a result, the obstacles that stand in the path of state reconstruction are exceptionally daunting.

Nor indeed is there any room in contemporary discourse on state collapse for the rise of armed refugee movements organized outside their homeland with the active support of external actors. Although the significance of the phenomenon transcends the cases at hand to include Somalia, Chad and Liberia, it is in Rwanda, Burundi, and Zaire that the collapse of state systems is most patently traceable to insurgencies born of refugee flows (8).

What needs to be stressed here is the potentially explosive mix arising from the involvement of conflict-generating refugee diasporas in electoral processes, a phenomenon made dramatically clear by the recent history of Rwanda. The decomposition of the state machinery on the eve of the genocide is traceable to the projection of electoral competition onto intra-Hutu power struggles involving alliances, real or presumed, with the Rwanda Patriotic Front (RPF), the external vehicle of Tutsi interests.

There are significant variations on this theme in Burundi and Zaire. In Burundi, the emergent Hutu-dominant state system born of the transitional elections of June 1993 was virtually blown to bits by the military coup of October 21, a move patently aimed at reversing the verdict of the polls. The assassination of the newly elected Hutu president, Melchior Ndadaye, did more than bring to a halt a five year transition to multiparty democracy; it created a deep split within Ndadaye's party (the Front Démocratique du Burundi -- Frodebu), between those who still wanted to give democracy a chance, and those who felt that recourse to force was their only option. Ambushed by extremists on both sides of the political spectrum, the Burundi state expired under the combined assault of Tutsi militias and army men, on the one hand, and Hutu rebels, on the other, the latter for the most part identified with the armed wing of the Frodebu in exile, now renamed the Conseil National pour la Défense de la Démocratie (CNDD).

In eastern Zaire the nationality issue -- who has the right to vote and who doesn't--lay at the heart of the "Kivu war" of 1993, pitting Banyarwanda (Hutu and Tutsi) against so-called "native" Zairois. More recently, the withdrawal of their citizenship rights was seen by many Tutsi as perfectly consistent with the threats voiced by the South Kivu authorities against "foreigners" and the exceptionally brutal "cleansing" operations directed against them by the Interhamwe and local units of the Forces Armées Zairoises (FAZ) in early 1996. After the wholesale slaughter of Tutsi civilians in Mokoto, in April 1996, thousands of them left the Masisi area to find refuge in Rwanda. The result of all this, as we now realize, was to create a set of mutually reinforcing conditions for a tactical alliance between the Banyamulenge and the APR, culminating in the shooting up of the refugee camps in October 1996 and the rise to power of Kabila.

The significance of such critical moments and events -- whether traceable to the surge of refugees across boundaries, the holding of transition elections or the intervention of the military -- points to yet another flaw in theories of state collapse: the very limited attention paid to triggering events in sharpening the edge of conflict, and accelerating the process of internal decomposition. A case in point is William Zartman's (9) casual dismissal of "turning points, warning signals, thresholds, or pressure spots," all of which offer important clues to an understanding of the disintegration of state systems in former Belgian Africa. Although there can be no quibbling over Zartman's characterization of state collapse as a "long-term degenerative disease", there is more to it than a "slippery slope" phenomenon. To these challenging thoughts we shall return in a moment. Let us, for the time being, take a closer look at the regional dimensions of state collapse.

REGIONAL FAULT LINES: The Kin-Country Syndrome

However wide of the mark the "clash of civilizations" may be in uncovering the roots of the Hutu-Tutsi conflict, what Huntington (10) refers to as the "kin country syndrome" is basic to an understanding of the process of escalation in the Great Lakes. Where ethnic fault-lines cut across national boundaries, conflict tends to spill-over from one arena to the next, transforming kin solidarities into a powerful vector of transnational violence. An action-reaction pattern sets in whereby victims in one setting become aggressors in the other. Such, in a nutshell, is the essence of the kin-rallying syndrome behind the escalation of violence in the region.

In such circumstances, as Huntington (11) reminds us, "conflict does not flow down from above, it bubbles up from below." At the heart of this bottom-up dynamic lies a phenomenon whose devastating effects are nowhere more dramatically revealed than in the three states under consideration -- the transformation of refugee-generating conflicts into conflict-generating refugees.

A critical aspect of regional escalation lies in the presence in each country of a large number of refugees, most of them with searing memories of the violence they experienced -- or inflicted -- in their homelands. Refugee flows can best be seen as the vehicles through which emotions are unleashed, ethnic ties manipulated, collective energies mobilized, and external support secured. What is at stake here is not simply the physical survival of human beings, but the political survival of specific ethnic communities. Whether as instruments in the hands of extremists for extracting assistance from humanitarian agencies, making deals with local authorities, or forging alliances with local kin groups -- or indeed as a political resource used by host governments or secondary level participants to further their foreign policy goals -- refugee movements, as one observer noted, are intensely political: "they create domestic instability, generate interstate tension and threaten international security"(12).

Much of the history of Hutu-Tutsi confrontations in Rwanda and Burundi is indeed reducible to the polarization of group identities that has accompanied the movement of refugee populations from one state to the other. Consider, for a moment, what happened in Burundi in the wake of the Rwanda revolution (1959-1962): of all the factors that have contributed to sharpen the edge of the Hutu-Tutsi conflict none has been more decisive than the flight into Burundi of some 60,000 Tutsi refugees from Rwanda in 1960-61, rendered homeless by Hutu-instigated violence, many mourning the death of relatives (13). By 1990 their number had grown to 180,000. Little wonder if some took an active part in the Burundi genocide of Hutu in 1972. But by then the shoe was on the other foot, with more than 50,000 Hutu from Burundi seeking asylum in neighboring states. Another major exodus of Hutu took place after the so-called "Ntega and Marangara incidents", in 1988, when in the wake of a local uprising thousands of Hutu fled to Rwanda to escape retribution from the army. Until the 1994 genocide in Rwanda, the most significant of such migrations took place immediately after Ndadaye's assassination, in October 1993, and the subsequent killing of thousands of Tutsi civilians by Hutu. As the all-Tutsi army proceeded to restore "peace and order" with its customary brutality, an estimated 300,000 Hutu poured across the boundary into Rwanda. As much as the devastating news of Ndadaye's death, their presence in Rwanda contributed in no small way to sharpen ethnic tensions. Although the evidence is lacking, there is every reason to believe that among the participants in the 1994 genocide were a fair number of Hutu from Burundi.

If domestic instability and interstate tensions are prominent features of the "kin country" syndrome in Rwanda and Burundi, the threats to regional security posed by the Kabiliste insurrection are equally clear. So also is the part played by violence-generating refugees in precipitating the crisis that paved the way for the insurrection.

To get our bearings on the significance of ethnic fault lines in North and South Kivu, and properly grasp the factors behind the abrupt reconfiguration of the social landscape of the area, we need to go back in time to the pre-genocide situation. A glance at the ethnic map of eastern Zaire reveals notable contrasts with what can be observed in Rwanda and Burundi. First and most obvious is the co-existence within the same provincial arenas of so-called "Banyarwanda" (Hutu and Tutsi) and a variety of ethnic communities indigenous to the region (Hunde, Nande, Banyanga, etc.). In 1993 the Banyarwanda were said to represent approximately half of the total population of 3.5 million in North Kivu. Of these, approximately 80 percent were Hutu and 20 per cent Tutsi (14). Furthermore, although neither Hutu nor Tutsi are a homogeneous lot, until recently the tendency among the "native tribes" has been to lump them together as "Banyarwanda," and to use the label as synonym for "foreign intruders." That a sizable number of them happened to be long-time residents of the area, that many were born in Zaire, or traced the origins of their families to pre-colonial migrations, seemingly made no difference.

Looked at in terms of clan and regional ties, and length of residence, there are of course marked differences among both Hutu and Tutsi. Some Hutu clans migrated into North Kivu long before the advent of colonial rule. The same is true of the Tutsi. Especially relevant in this regard is the case of the Banyamulenge, a Tutsi sub-group of South Kivu, numbering in the tens of thousands (15). The label, in its pre-genocide connotation, suggests a sense of identity derived from the locality where they first settled, that is "the people of Mulenge." The Banyamulenge are only one of a number of Banyarwanda communities that might be referred to as "the early settlers," people who migrated to North and South Kivu in the early or mid-nineteenth century, if not before; another category are the so-called "transplantes" i.e., the thousands of workers (mostly Hutu) who were brought to the Kivu at the request of the colonial administration to work on tea and coffee plantations; a third group are the mostly Tutsi refugees who left Rwanda in the early sixties during the Hutu revolution, numbering anywhere from 30,000 to 50,000. For all these differences, the Banyarwanda never ceased to be seen by the soi-disant Zairois otherwise than as foreigners, and therefore disqualified to claim citizenship rights. No other issue has had a more decisive impact on the rise of a collective self-awareness among the Banyarwanda than the withdrawal of their citizenship rights by the 1981 Nationality Act (16). After the RPF invasion of Rwanda in 1990 incipient Hutu-Tutsi tensions began to emerge, but these were consistently kept in check by their common awareness of anti-Banyarwanda sentiment among "native tribes."

With the massive outpouring of Hutu refugees and Interhamwe from Rwanda in 1994 the Banyarwanda frame of reference quickly dissolved into a rigid Hutu-Tutsi dichotomy. An instant sea-change occurred both in the perceptions that the Banyarwanda had of each other, and in the images that Hutu and Tutsi projected of themselves in the social milieu of eastern Zaire. The "kin-country syndrome" asserted itself with a vengeance, driving Hutu and Tutsi, irrespective of other distinctions, into opposing camps. Meanwhile, the communities indigenous to North Kivu began to cast about for tactical alliances. While "native" Hutu joined hands with the Interhamwe, FAR and Hunde elements against the "native" Tutsi, the latter responded by casting their lot with the RPA (Rwandan Patriotic Army), but not before thousands of them had been slaughtered by Interhamwe and FAZ elements in the Spring of 1996.

Before long the Banyamulenge label was freely used as an all-encompassing identity marker by all Tutsi in Zaire, irrespective of regional ties or length of residence. Behind this curious case of ethno-genesis lies a clear political objective: to openly proclaim their Zairian roots and their full rights to Zairian citizenship. At the semantic level at least the term Banyamulenge has settled once and for all the nationality issue. This is not the place for a full-scale discussion of how the fault-line war in eastern Zaire escalated into a regional conflict, with Rwanda, Uganda, and to a lesser extent Burundi, as secondary participants, and Angola and the Sudan as tertiary parties. Suffice it to note that the collapse of the Zairian state could not have happened so swiftly, and with so little resistance from the FAZ, unless the Mobutist state had already shown alarming signs of decomposition.

STATE COLLAPSE: The Longue Durée Dimension

State decay does not happen overnight. It is a long-term process, which brings to mind Braudel's 'long duration' (longue durée) dimension. Although the Braudelian facet of state collapse applies to all three countries, the cases of Rwanda and Burundi reveal contextual specificities that mark them off sharply from their neighbor to the west.

In bi-ethnic, vertically structured social arenas, exclusionary policies are a major source of erosion of state legitimacy. With the benefit of hindsight, one can better appreciate the long-term implications of such policies for the Rwanda state. The more or less systematic exclusion of Tutsi residents in Rwanda from meaningful political participation (beyond a quota system that left few illusions in the minds of its presumptive beneficiaries), along with the refusal of virtually every government to allow the Tutsi population in exile to return to their homeland, made the Rwanda state doubly vulnerable. It created a deep and lasting sense of alienation among the resident Tutsi population--in time making them highly receptive to the appeals of their kin-group in exile -- while providing the exiles with justification for the 1990 invasion.

There is an obvious parallel between the exclusion of Tutsi in Rwanda and of Hutu in Burundi, except for the fact that the latter formed the overwhelming majority of the population. Moreover, the exclusion of Hutu in Burundi occurred more gradually, at a later stage, and did not reach completion until the 1972 genocide, when anywhere from 100,000 to 200,000 Hutu were massacred in the wake of an abortive insurrection, and tens of thousands forced into exile. Again, at no time did they benefit from anything comparable to the massive military and logistical support extended by Uganda to the RPF. What needs to be stressed is that in both Rwanda and Burundi, ethnic exclusion resulted in the birth of an ethnocratic state made all the more vulnerable by the rapid shrinking of its power base (of which more later).

As the recent history of the Banyarwanda in eastern Zaire cruelly shows, the denial of their citizenship rights was certainly a factor in the concatenation of events leading to the unraveling of the Mobutist state. Unlike what happened in Rwanda and Burundi, however, ethnic exclusion did not hold the same implications. As long as it involved peripheral communities -- the Banyarwanda in Kivu, the Luba in Shaba, the Bakongo in the Lower Congo -- the threats to the center seemed manageable. Although very much part of Mobutu's manipulative tactics, ethnic exclusion has never been a systematic, guiding principle of Zairian policies. If anything, the striking policy inconsistencies surrounding the nationality question in the Kivu are better seen as the symptom rather than the cause of the incoherence of the Mobutist state.

The threat of Malthusian trends is the second major source of state erosion that needs to be stressed, in part because it gave justification to the first. Long before they reached the edge of the precipice both states were faced with demographic pressures which neither had the capacity to contain. Rwanda and Burundi claim the highest population densities in the continent. At the turn of the century each state had a population of roughly 1.5 million. Today their combined populations are close to 15 million. In comparative terms this is as if one third of Zaire's population were enclosed in a space one-sixtieth of its size. While raising serious questions about the long-term viability of their state systems, the ever-expanding size of their population retrospectively explains the past reluctance of their governments to accept the return of large numbers of refugees. Moreover, it focuses attention on land shortage as a major ingredient of ethnic strife (particularly in Rwanda, but also in eastern Zaire) and the inability of the state to develop policies designed to reallocate land effectively and equitably, the reason in part being that land eventually became a key resource in the arsenal of the state to build faithful clienteles among privileged social groupings.

We touch here on a third aspect of the long-term processes of state decay, with relevance to all three cases at hand: the shrinking of the political bases of state authority. Nothing is more revealing of the weakening of the Rwanda state under Habyalimana than the steady erosion of its power base. The shift of power accomplished by the revolution, from Tutsi to Hutu, uncovered deep regional fractures among Hutu, northerners vs. southerners, Bashiru vs. Bagoye, etc. As the regional struggles over patronage intensified, the point was reached where power and authority tended to gravitate increasingly around the presidential household and his immediate family, the so-called akazu ("small house" in Kinyarwanda). At the time of the RPF invasion, on October 1, 1990, the Rwanda state was little more than a caricature of the neo-patrimonial polity.

Much the same contraction of the political arena can be seen in Burundi. From 1972 onwards, power became the exclusive privilege of Tutsi-Hima elements from the south (Bururi), a situation that came to reflect the dominant position of Tutsi-Hima officers in the military. If nothing comparable to the akazu phenomenon characterized the Burundi situation, there can be little doubt about the inherent fragility of a state system where key decisions are made by a handful of army officers, and where the army itself is subject to ceaseless internecine struggles. The overall implications are well summed up by Rothchild and Groth (17):

Because state institutions are fragile and lacking in effectiveness and legitimacy, they are a poor vantage point to mediate the struggle between competing groups. Unable to channel participation along predetermined lines, the overloaded state becomes isolated and aloof from society, unable to structure the relations between social interests or between these interests and itself.

Nowhere is this loss of legitimacy and growing isolation more palpably evident than in Zaire. For a quarter of a century the Mobutist state was able to compensate for its lack of internal legitimacy by drawing huge dividends from its international status as the staunchest ally of the United States in Africa. The end of the Cold War could not but sharply increase its international isolation and legitimacy deficit; bartering its anti-communist credentials for external assistance was no longer a feasible option. Just as Mobutu owed his rise to power to the incidence of East-West rivalries in the continent, in the last analysis the collapse of the Zairian state must be seen as a casualty of the Cold War's end.

Intimations of the mortality of the Mobutist state were felt long before its downfall. Its multiple afflictions have been diagnosed in considerable detail by Turner and Young (18). Some are rooted in the cumulative effect of economic and financial constraints ranging from the plummeting of copper prices in the 1970s and the ineptitudes of "Zairianisation", to a growing debt burden and a widening gulf between a soaring supply of money and the availability of basic commodities, leading to runaway inflation. Others are clearly traceable to Mobutu's own neo-patrimonial style, which conjures up mixed images--Bula Matari working in tandem with the Medellin cartel or Cosa Nostra. The result has been a process of political involution centered around a handful of rent-seeking cronies, leading to what Crawford Young pithily describes as "self-cannibalization": "the state consumes itself to live for another day". "The decay of the public realm," he goes on to note (19),

is marked by a cumulative deflation of the state apparatus in terms of its competence, probity and credibility. Institutions of rule lose their capacity to translate public resources into sustenance of infrastructures or valued amenities. A pervasive venality surrounds most public transactions. As a consequence, the subject comes to experience rule as simple predation; the aura of the state as powerful and nurturant protector vanishes.

While the image of the state as protector receded, that of the state as predator came increasingly into focus. To compensate for the unpaid salaries of his troops, Mobutu in effect gave them a blank check to ransom and loot. The privileged ethno-regional clienteles built around the Ngbandi-dominated Division Spéciale Présidentielle (DSP) only reinforced the disaffection of the troops, whose principal source of livelihood was plunder and theft. The phenomenon was already patently clear in the early 1990s if not earlier, and became all more threatening during the Kabiliste insurrection. For the majority of the troops sent out to crush the rebellion the purpose of their assignment was not to defeat the enemy, but to take maximum advantage of the situation to engage in one "pillage" after another. As the tide began to turn many ended up selling their weapons to potential rebels, or joining their ranks.

In the catalog of forces that conspired to produce the ultimate collapse of the state in Rwanda, Burundi and Zaire, some are specific to their history and socio-ethnic configurations, others to the complex pattern of interaction arising from the "kin-country" syndrome. Yet in all three states emerges a common denominator: the extension of the dysfunctions of the state to their militaries. The privatization of the Zairian army and its propensity to instigate civil violence for purposes of personal profit, the active participation of the FAR in the Rwanda genocide, and the murderous intervention of the Burundi army to block the transition to democracy--all are reflective of the declining capacity of the state to control its instruments of coercion. From all evidence, the Weberian definition of the state may well provide a more useful thread for identifying the roots of its disintegration than some of the more fashionable extant taxonomies (20).

TRIGGERS AND THRESHOLDS

In an otherwise inspired essay, Zartman (21) makes surprisingly short shrift of the multiplier effect of decisive events on processes of state collapse. His use of metaphors is revealing: "What is notable in these scenarios (of state collapse) is the absence of clear turning points, warning signals, thresholds, or pressure spots. . . The slippery slope, the descending spiral,and the downward trend are the mark of state collapse rather than deadlines and triggers."

The least that can be said of this curiously ahistorical construction is that it is difficult to reconcile with the evidence at hand--not unlike trying to explain the fall of the French monarchy without reference to the seizure of the Bastille, the Tennis Court Oath, or the Flight to Varenne. Triggers are not to be dismissed lightly, least of all when directed against entire ethnic communities--or when perceived as such. From this vantage the October 1, 1990 invasion of Rwanda by the Rwanda Patriotic Front (RPF) can only be seen as a watershed. Against heavy odds, almost four years later, the RPF was able to claim victory, but at an appalling cost.

Immediately seen by the Habyalimana government as evidence of a Tutsi conspiracy, the invasion was the signal for arresting tens of thousands of Tutsi civilians, and not a few Hutu, throughout the country. In the capital city an estimated 100,000 people were brought at gun point into the national stadium for questioning. Only months later were they finally allowed to return home. By then, however, ethnic violence had crossed new thresholds of intensity. In January and February 1991, in response to a daring RPF raid on the Ruhengeri jail, local Hutu militias massacred hundreds of Bagogwe pastoralists (a Tutsi subgroup). This was only the first in a series of anti-Tutsi pogroms culminating in March 1992 with the cold-blooded massacre of thousands of Tutsi civilians in the Bugesera region. Predictably, the Tutsi invaders for their part showed little restraint in dealing with Hutu civilians in the "liberated" areas. Tens of thousands are said to have been slaughtered by advancing RPF troops. By January 1993, an estimated one million Hutu civilians were forced into camps throughout the country. Only if one remembers their desperate condition--having lost all their possessions and sometimes their relatives--and the depth of anti-Tutsi sentiment, can one understand why so many of them ended up supplying the bulk of Habyalimana's "genocidaires."

With the assassination of Burundi's first elected Hutu president, on October 21, 1993, another threshold was crossed in the rising tide of ethnic hatreds. In Rwanda, Ndadaye's murder carried enormous symbolic significance. The message, in its devastating simplicity, came through loud and clear: "You just cannot trust the Tutsi!" Relayed in the countryside by thousands of panic-stricken refugees, the warning also found an immediate echo in the media, including the racist Radio Mille Collines. Not only did the news of Ndadaye's death virtually destroy the fragile consensus built around the Arusha accords; from then on the Hutu militancy was clearly on the ascendant, transforming the state into a battle ground between moderates and radicals.

The third, most lethal trigger was the two surface-to-air missiles that brought down President Habyalimana's plane over Kigali, on April 6, 1994, killing, along with Habyalimana and his closest aides, President Cyprien Ntaryamira of Burundi. To this day the identity of the men who fired the missiles remains a mystery (22). Most Hutu, however, immediately detected the hand of the RPF behind the dastardly deed. The killings began within hours.

As the Rwanda state finally collapsed, awash in a sea of blood, its counterpart in Burundi was barely able to sustain the backlash of the genocide next door. As in a game of mirrors, reflecting symmetrical images, the Rwanda genocide quickly entered the consciousness of the Tutsi community of Burundi in the form of a self-fulfilling fantasy, giving retrospective justification to Ndadaye's assassination, thus becoming part and parcel of what Paul Veyne calls "l'imagination constituante"(23). The killings in Rwanda were not just an omen of what could happen to the Tutsi in Burundi; genocide had already happened. What some erroneously viewed as a spontaneous explosion of violence after Ndadaye's death, was nothing less than genocide. Indeed, it was part and parcel of Ndadaye's dastardly plans (which is why he had to be killed). It mattered little that (a) Burundi was the scene of the first genocide recorded on the continent, and (b) that it was a genocide of Hutu by Tutsi. The only point that mattered for the more militant elements of the Tutsi community was to capture the moral high ground by holding the majority of Frodebu politicians collectively responsible for the genocide of their kin group. The drawing of a wholly arbitrary line of demarcation around the "genocidaires" also meant drawing of state boundaries. As parliament ceased to function, and as governments were made and unmade by pressures from the street, orchestrated by Tutsi militias, what was left of the state fell into the hands of the Tutsi-dominated army, which to this day remains the principal arbiter of conflict.

Space limitations do not permit more than the briefest reference to the equally devastating impact of the Rwanda genocide in eastern Zaire. Here again it is the catalytic effect on group identities that needs to be stressed. With the massive surge of Hutu refugees into the Goma area, followed in early 1996 by countless atrocities committed by Interhamwe against local Tutsi, collective identities quickly sorted themselves out into rival communities. Once allies and victims in their fight for Zairian citizenship, they suddenly turned against each other with appalling ferocity. Long before eastern Zaire became the launching pad of Kabila's revolution, by courtesy of Vice-President Kagame, North Kivu had been the scene of a hideous slaughter of Banyarwanda in 1993 (24). As much as anything else about the Zairian state, its inability to come to grips with the nationality question, and with the roots of the 1993 carnage, revealed the extent of its paralysis.
If the foregoing does illuminate certain critical dimensions of state collapse, it also hints at some of the more intractable problems that lie ahead on the road to reconstruction.

PROSPECTS FOR RECONSTRUCTION

After the descent into hell, comes the purgatory of national reconstruction. Each of the three states under consideration is entering this somewhat opaque halfway house with uneven indulgences, and at different speeds. Although Rwanda has already made commendable progress, Congo/Zaire has barely crossed the threshold of redemption. Only in Burundi is the state fated to remain in limbo for the foreseeable future. Clearly, given the extreme fluidity of the regional context, any attempt to assess the prospects for reconstruction must be highly speculative. By way of a starting point, two remarks are worth noting. Firstly, only in Rwanda has something resembling a state system re-emerged from the chaos of genocide; in Burundi and Zaire, on the other hand, the nearest thing to the state is what Misha Glenny, in his classic work on the fall of Yugoslavia, calls "the parastate," i.e. "the mutant offspring of an expiring failed state. . . boasting certain essential attributes of a normal state but grotesquely lacking in others"(25). Because of its much greater degree of "stateness," Rwanda was able to play a critical role in destabilizing Burundi's "genocidaires" in exile and in "facilitating" the Zairian transition. There is every reason to believe that Rwanda will expect substantial dividends in return, primarily in the form of continued influence in Kinshasa and Bujumbura.

Secondly, just as Habyalimana's Rwanda was the model polity which many Hutu would have liked to transpose into Burundi, many are the Tutsi politicians in Bujumbura, including some that might be described as "moderates," who today look to Kagame's Rwanda for inspiration. Kagame's Rwanda has all the earmarks of an ethnocracy, but with just enough power sharing at the top to enlist a measure of Hutu collaboration. Thus, if the substance of power is to remain in Tutsi hands, via the military, no effort should be spared to give Hutu elements willing to collaborate a prudent share of participation in decision-making.

Even the most cursory glance at the pattern of reconstruction in Rwanda cannot fail to notice the characteristic traits of a military ethnocracy. The emergent polity is one in which the 45,000 strong, all-Tutsi army provides the critical underpinning of the formal government institutions. Key decisions are made by Kagame and his trusted lieutenants, flowing from the top down. The appointed parliament is little more than a fig leaf that barely conceals the dominant position of the RPF. The civil service, the judiciary, the economy, the schools and university are all under Tutsi control (26). The closest thing to a constitution are the Arusha accords of August 1993. Although intended to provide the basis for an all-embracing power-sharing formula, extending to the armed forces, the accords have been consistently "adapted" and manipulated to serve the policy goals of the regime. Efforts to rebuild the judicial system are proceeding with less than optimal results. While the trials of genocide suspects are said to be relatively fair, there are still some 100,000 Hutu languishing in Rwanda's jails. Quite aside from the fact that approximately half of the detainees are said to have been incarcerated for reasons having little to do with their presumed involvement in genocide, but rather as the quickest way for their neighbors to grab their property, the conditions in which most prisoners are being held can only be described as inhuman.

With ethnic violence picking up momentum in the north -- largely as a result of armed raids by repatriated Interhamwe and ex-FAR, inevitably followed by a devastating retribution in kind by the APR -- the prospects for enlarging the ethnic base of the state appear extremely remote. The implications transcend the Rwanda arena. In Burundi, where the radical fringe of the Tutsi community remains extremely sensitive to the lessons of Rwanda, there are signs that the current efforts at mediation will be violently rejected by certain units of the army and the militias, with the Bagazistes trying to draw maximum advantage of the situation to further advance their ethnocratic claims. In Zaire/Congo, any move designed to curtail the influence of Rwanda's "near abroad" will probably be met with stiff resistance from Kagame, and indeed from the Banyamulenge currently in charge of the security and the local administration in North and South Kivu. Clearly, issues of ethnicity will continue to set the key parameters for reconstruction in all three states. Which brings us to a brief consideration of the impending avatars of the Kabiliste take-over in the DCR.

The difficulties facing Kabila are inscribed, in part, in the circumstances of his astoundingly rapid military victory. Three factors are particularly worth noting: (a) the critical role played by Banyamulenge elements trained in Rwanda in cementing the politico-military armature of the rebellion; (b) the all-pervasive, overwhelming anti-Mobutu sentiment that infused the civil society, and caused hundreds of thousands of Zairians to cast their lot with the rebels, long before they even came into view; (c) the widespread assumption among anti-Mobutist forces that "liberation" means a swift transition from dictatorship to democracy.

For Kabila to break out of the ethnic enclave in which he is now entrapped is a key priority if his reconstruction project is to gain legitimacy. The task will not be easy. With the rapid advance of the rebellion into the interior, a growing number of Zairians from almost every province were brought into the armed forces of the Alliance, thus diluting the salience of the Banyamulenge presence in the military, but not to the point of allaying fears of "foreign domination." Banyamulenge form the hard core of Kabila's troops; many occupy key positions in the administrative machinery of North and South Kivu. Anti-Banyamulenge feelings run high in both provinces, particularly among Hunde and Babembe. Despite efforts to incorporate Bashi elements in the provincial power structure their loyalty is open to question. The same is true of the Baluba/Kasai, who bitterly resent the exclusion of their "favorite son" (Etienne Tshisekedi) from the ruling government coalition. Kabila is thus faced with a Hobsons's choice: failure to meet the imperative of a broadly based coalition, meaning also a genuine effort to scale down the influence of the Banyamulenge in the Kivus and elsewhere, can only lead to a loss of legitimacy; turning against the architects of his victory against Mobutu, on the other hand, would be tantamount to political suicide.

Kabila must bear the unanticipated costs of a military conquest that quickly outpaced his capacity to put in place a viable administration. The abrupt collapse of the Mobutist state has created a political vacuum that has yet to be filled. The most notable exception is North and South Kivu. Even so, the picture conveyed to outside observers is one of considerable improvisation, with little attention paid to the potential support that could be derived from the civil society. If the situation in the Kivus is any index, many are the civil society organizations (CSO) that could have provided the social ballast needed to reconstruct the new polity, but so far their place in the new dispensation appears extremely nebulous. Many have been torn apart by ethnic rivalries born of the rebellion; some were simply dismantled, while others were brought under the tight control of Alliance cadres. Fear that the CSOs could transform themselves into "contre-pouvoirs" is all pervasive.

In the absence of a civil society capable of providing effective linkages with the state, the day-to-day tasks of administering the liberated territories have been entrusted to the Commissaires de Zone. Although in many instances local incumbents were allowed to remain in office, there can be little doubt as to where power lies -- in the hands of the Commissaires, acting hand in hand with local units of the armed forces. Though decidedly more disciplined than the FAZ, at times the enforcement of discipline on the civilian population is ominously reminiscent of the Bula Matari scenarios. "The attitude of the army," according to a first-hand witness, "is designed to bring back a taste of civic mores ('kuleta morale'), with an introduction of the chicotte. . . What is unacceptable by any modern standard of justice is the fact that whip lashing (on the legs) takes place on the spot, lying down face on the ground, by the same people who observe the alleged misdemeanor" (27). If this testimony -- eerily evocative of the most somber of Tshibumba's paintings (most notably "Colonie Belge") -- is any indication, recourse to force could figure prominently in Kabila's strategies of state reconstruction. This impression is strongly reinforced by the deliberate, wanton killings of tens of thousands of Hutu refugees at the hands of Alliance troops, prompting the EU Head of Humanitarian Affairs Commission, Emma Bonino, to describe the killing grounds of eastern Zaire as a "slaughter house."

As the foregoing plainly suggests, the focus of Kabila's efforts at reconstruction is less on democracy than on the creation of spaces for discipline and moralization. In view of his own background as an Afro-Marxist guerrilla fighter during the 1964 rebellion, and his subsequent checkered career, it is easy to see why the virtues of the Civic Culture should only have a limited appeal to his projet de société. His early exposure to Marxism has left an ideological legacy that points to a systematic effort at the re-socialization of Zairian society. In the Kivu this finds expression in the ideological seminars conducted by the secrétaires généraux charges de la coordination, in which the emphasis is on a class analysis of Zairian society; in terms of organization the aim is to reach out to the grassroots through the local cells (Chembe Chembe) set up to assist the administration; ultimately it is for centrally appointed village officials to filter and sanction the "general will" expected to emanate from the rural masses. It is evidently far too early to draw definitive conclusions from the situation observable in the Kivus; the ambivalence of Kabila's project (28) is well captured in the description offered by a recent visitor to eastern Zaire:

Whereas certain statements of the AFDL (Alliance des Forces D^émocratiques pour la Lib^ération du Congo) (interviews with Kabila, etc.) stress a liberal approach, the ideological courses to which all functionaries are subjected rather stress a Marxist approach (class struggle, national bourgeoisie, etc.). An article in the Rwandese official newspaper La Nouvelle Releve (no. 339 of 7/4/97) gives yet another viewpoint, which centers around "a new social order in a rural environment" comparable to the ideas of Guinea-Bissau's Cabral for an African Socialism. Summarized in very short order, it states that organizing the rural society is first and foremost in solving the country's problems. This can be concretized in three points: changing political structures in favor of small producers, changing the mode of production, and increasing the productivity. To enable this accelerated development, a combination of state efforts (parastatals) and private efforts (via cooperatives) will be necessary.

It will be some time before the contours of the new state can be discerned with any degree of precision -- beyond what few tentative conclusions one may draw from the profile of the newly appointed government. Whether the expectations of the Zairian/Congolese masses can be met in time to prevent a major political backlash remains unclear. What is beyond doubt is that more will be required than a formal commitment to democracy, or Afro-Marxism, to reinvigorate private enterprise, restart the production of the industrial and manufacturing sectors, get the parastatals back on the rails, and restore the infrastructures. From Mobutu's kleptocratic rule Kabila has inherited a devastated economy, a society driven by ethno-regional enmities, plagued by deep poverty and shocking social inequalities, a country which, as one rebel radio broadcast noted, "has been crushed to a pulp." Rebuilding the DCR on the ashes of the Zairian state promises to be a Herculean task. Whether Kabila proves equal to the challenge remains to be seen.

ENDNOTES

This paper was prepared for presentation at the XVIIth World Congress of the International Political Science Association, 17-21 August 1997, Seoul, South Korea, Copyright IPSA 1997. None of the views set forth in this paper are to be attributed to USAID or any other agency of the US government; I claim full responsibility for all errors of fact or interpretation.

(1) The question of how many refugees returned to Rwanda, and how many stayed behind, is a highly controversial issue. By and large pro-Rwanda analysts tend to greatly inflate the number of returnees (up to 700,000) and scale down the number of those who stayed in Zaire. Typical of this tendency to manipulate statistics is the following statement from a US official in Kigali: "Half a million refugees did not remain behind, but only about 100 to 200,000. . . . The UNHCR and NGOs grossly over-counted the refugees by mistake or probably more likely for their own motives -- continued western funding" (Personal communication). According to UNHCR figures, the total refugee population, before October 1996, was in the order of 1.1 million distributed among some fifteen camps from Mugunga in the north to Uvira in the south. Some 316,000 were reported in the camps around Bukavu, 715,000 around Goma, and 180,000 around Uvira, of whom 117,000 were Barundi. Assuming that as many as 700,000 returned to Rwanda, which has yet to be confirmed, this still leaves some 400,000 "unaccounted for." For an excellent survey of the refugee situation in eastern Zaire see the special issue of Dialogue ("Les Oublies de l'Afrique des Grands Lacs"), No. 196 (February 1997), as well as Johann Pottier's outstanding discussion, "The 'Self' in Self-Repatriation: Closing Down Mugunga Camp (Eastern Zaire)", forthcoming, in Richard Black and Khalid Koser, eds., The End of the Refugee Cycle? Refugee Repatriation and Reconstruction, Bergahm Books.

(2) Samuel Huntington, "The Clash of Civilizations and the Remaking of World Order." Foreign Affairs, V. 72, N. 3, Summer 1993, p22(28). The Clash of Civilizations and the Remaking of World Order. New York: Simon and Schuster, 1996.

(3) Ibid. , p. 28.

(4) For further elaboration on this theme, see René Lemarchand, Burundi: Ethnic Conflict and Genocide. New York: Cambridge University Press and Woodrow Wilson Center Press, 1995.

(5) For an illuminating discussion of the constructivist and other models of ethnicity, see Crawford M. Young, "The Dialectics of Cultural Pluralism: Concept and Reality", in Crawford M. Young, ed., The Rising Tide of Cultural Pluralism. Madison: University of Wisconsin Press, 1993, p. 21. On the role of intellectuals in fashioning the constructivist face of ethnicity in Rwanda and elsewhere, see Michael Chege, "Africa's Murderous Professors", The National Interest, Winter 1996-97.

(6) Thomas Carothers, "Which Democracy Should We Export", Harper's Magazine, September 1996, p. 19.

(7) Alex De Waal, "The Genocidal State", Times Literary Supplement, July 1994, n. 4761, p.3. For an illuminating background discussion of the Rwanda genocide, see David Rieff, "An Age of Genocide: The Far-Reaching Lessons of Rwanda", The New Republic, January 26, 1996.

(8) On the genesis of refugee flows, see Myron Weiner, "Bad Neighbors, Bad Neighborhoods: An Enquiry into the Causes of Refugee Flows", International Security, vol. 21, No. 1 Summer 1996, pp. 5-42.

(9) William Zartman, "Introduction: Posing the Problem of State Collapse", in William Zartman, ed., Collapsed States. Boulder: Lynne Rienner, 1995.

(10) Huntington, op. cit., p. 272.

(11) Ibid. , loc. cit.

(12) G. Loescher, "Refugee Movements and International Security", Adelphi Paper No. 268, International Institute of Strategic Studies, London 1992, quoted in Liisa Malkki, "Refugees and Exile: From 'Refugee Studies' to the Rational Order of Things", Annual Review of Anthropology, Vol. 24, 1995, pp. 504.

(13) For the historical background, see René Lemarchand, Rwanda and Burundi. New York and London: Pall Mall Press and Frederick Praeger, 1970.

(14) See Filip Reyntjens and S. Marysse. eds., Conflits au Kivu: Antecedents et Enjeux. University of Antwerp: Center for the Study of the Great Lakes Region of Africa, Dec. 1996.

(15) The size of the Banyamulenge population in South Kivu remains an enigma; the figures cited vary from 35,000 to 200,000 (see Reyntjens and Marysse, op. cit.). An educated guess would suggest at least 60,000, possibly more; the figure of 200,000, cited by International Alert and other NGOs, seems grossly exaggerated. Part of the confusion may stem from the failure to take into account the changing definition of Banyamulenge: after the attacks on the refugee camps, the term came to designate all Tutsi, regardless of whether they came from North or South Kivu, or their length of residence in Zaire. See the outstanding analysis in James Fairhead, "Demographic Issues in the Great Lakes Region", Save the Children Conference on Practical Approaches to the Crises of the Great Lakes, Sunridge Park, London, 24-26 March 1997. "Statistics from 1991 in Rwanda," Fairhead writes, "suggested a catastrophic scenario within the next 25 years." Present populations average 600 people per usable ha, with average rural population density set to increase to c. 1000-1500 inhabitants per sq. km in 2015, with average farms of less than 1 ha on which 8-12 people must live." Population projections for Burundi and eastern Zaire, he adds, suggest equally problematic scenarios. See Johann Pottier, "Social Dynamics of Land and Land Reform in Rwanda: Past, Present and Future", SOAS, University of London. Typescript, April 1997.

(16) The 1981 Nationality Act limited Zairian citizenship only to those persons able to show they had an ancestor belonging "to one of the tribes established in Zaire since 1865," thus repealing the 1972 law granting citizenship rights to all Banyarwanda established in Zaire before January 1950. Not only did the 1982 Act disqualify those tens of thousands of Banyarwanda and Barundi who came into Zaire at the request of the Belgian authorities, along with Tutsi elements who fled the Rwanda revolution; even more exasperating was the absurdity of a piece of legislation that made it virtually impossible for anyone to comply with provisions given that (a) the boundaries of Zaire in 1885 had yet to be fixed, and (b) proof of a pre-1885 ancestry is impossible to establish in juridical terms. For further details on the nationality issue, see Jean-Claude Willame, Banyarwanda et Banyamulenge; Violences Ethniques et Gestion de L'Identitaire au Kivu Paris: L'Hartmattan, 1997)

(17) Donald Rothchild and Alexander J. Groth, "Pathological Dimensions of Domestic and International Ethnicity", Political Science Quarterly, Vol. 110, No. 1, 1995, p. 74.

(18) Crawford M. Young and Thomas Turner, The Rise and Decline of the Zairian State. Madison: University of Wisconsin Press, 1988.

(19) Crawford M. Young, "Reflections on State Decline and Societal Change in Zaire" Typescript, January 1997, p. 2.

(20) See, for example, Jean-Germain Gros, "Towards a taxonomy of failed states in the New World Order: decaying Somalia, Liberia, Rwanda and Burundi", Third World Quarterly, Vol. 17, No. 3 1996, pp. 455-471, where the author valiantly wrestles with five categories of failed polities: "anarchic states", "phantom or mirage states", "anaemic states", "captured states", and "states that failed in vitro (they are called aborted states)."

(21) Zartman, op. cit., p. 9. The other side of this methodological coin is an overly functionalist approach to the subject of state collapse. "Why do states collapse," asks Zartman. "Because they can no longer perform the functions required for them to pass as states" (p. 5). What would the author reply to a doctor who would explain the death of his patient by gravely announcing that he/she could no longer perform the functions required to stay alive?

(22) For an instructive, although inconclusive, effort to solve the mystery, see Filip Reyntjens, Rwanda: Trois jours qui ont fait basculer l'histoire Brussels: Institut Africain/CEDAF, and L'Harmattan: Paris, 1995.

(23) Paul Veyne, Les Grecs ont-ils cru à leurs mythes? Essai sur l'imagination constituante Paris: Le Seuil, 1983. This myth has been largely endorsed by the UN Commission in charge of investigating the circumstances of Ndadaye's assassination.

(24) For an excellent first-hand account, see Davis Orr, "Kivu Province Becomes a War Zone", Focus on Africa, Vol. 4, No. 4 October-December 1993, pp. 5-8.

(25) Michael Glenny, The Fall of Yugoslavia London: Penguin Books, 1996, p. 263.

(26) According to a reliable source, out of a total of approximately 5,200 students registered at the National University in Butare, 5,000 are Tutsi and 200 Hutu.

(27) Anon., "Tupemata morale: Report of a Field Trip to South-Kivu", May 1-8,1997, p. 3.

(28) Ibid., p. 2.