Collapse: How Societies Choose to Fail or Succeed

by Jared Diamond

  In recent decades, Rwanda and neighboring Burundi have become synonymous in our minds with two things: high population, and genocide (Plate 21). They are the two most densely populated countries in Africa, and among the most densely populated in the world: Rwanda’s average population density is triple even that of Africa’s third most densely populated country (Nigeria), and 10 times that of neighboring Tanzania. Genocide in Rwanda produced the third largest body count among the world’s genocides since 1950, topped only by the killings of the 1970s in Cambodia and of 1971 in Bangladesh (at the time East Pakistan). Because Rwanda’s total population is 10 times smaller than that of Bangladesh, the scale of Rwanda’s genocide, measured in proportion to the total population killed, far exceeds that of Bangladesh and stands second only to Cambodia’s. Burundi’s genocide was on a smaller scale than Rwanda’s, yielding “only” a few hundred thousand victims. That still suffices to place Burundi seventh in the world since 1950 in its number of victims of genocide, and tied for fourth place in proportion of the population killed.

  We have come to associate genocide in Rwanda and Burundi with ethnic violence. Before we can understand what else besides ethnic violence was also involved, we need to begin with some background on the genocide’s course, the history leading up to it, and their usual interpretation that I shall now sketch, which runs as follows. (I shall mention later some respects in which this usual interpretation is wrong, incomplete, or oversimplified.) The populations of both countries consist of only two major groups, called the Hutu (originally about 85% of the population) and the Tutsi (about 15%). To a considerable degree, the two groups traditionally had filled different economic roles, the Hutu being principally farmers, the Tutsi pastoralists. It is often stated that the two groups look different, Hutu being on the average shorter, stockier, darker, flat-nosed, thick-lipped, and square-jawed, while Tutsi are taller, more slender, paler-skinned, thin-lipped, and narrow-chinned. The Hutu are usually assumed to have settled Rwanda and Burundi first, from the south and west, while the Tutsi are a Nilotic people who are assumed to have arrived later from the north and east and who established themselves as overlords over the Hutu. When German (1897) and then Belgian (1916) colonial governments took over, they found it expedient to govern through Tutsi intermediaries, whom they considered racially superior to Hutu because of the Tutsi’s paler skins and supposedly more European or “Hamitic” appearance. In the 1930s the Belgians required everybody to start carrying an identity card classifying themselves as Hutu or Tutsi, thereby markedly increasing the ethnic distinction that had already existed.

  Independence came to both countries in 1962. As independence approached, Hutu in both countries began struggling to overthrow Tutsi domination and to replace it with Hutu domination. Small incidents of violence escalated into spirals of killings of Tutsi by Hutu and of Hutu by Tutsi. The outcome in Burundi was that the Tutsi succeeded in retaining their domination, after Hutu rebellions in 1965 and 1970-72 followed by Tutsi killings of a few hundred thousand Hutu. (There is inevitably much uncertainty about this estimated number and many of the following numbers of deaths and exiles.) In Rwanda, however, the Hutu gained the upper hand and killed 20,000 (or perhaps only 10,000?) Tutsi in 1963. Over the course of the next two decades up to a million Rwandans, especially Tutsi, fled into exile in neighboring countries, from which they periodically attempted to invade Rwanda, resulting in further retaliatory killings of Tutsi by Hutu, until in 1973 the Hutu general Habyarimana staged a coup against the previous Hutu-dominated government and decided to leave the Tutsi in peace.

  Under Habyarimana, Rwanda prospered for 15 years and became a favorite recipient of foreign aid from overseas donors, who could point to a peaceful country with improving health, education, and economic indicators. Unfortunately, Rwanda’s economic improvement became halted by drought and accumulating environmental problems (especially deforestation, soil erosion, and soil fertility losses), capped in 1989 by a steep decline in world prices for Rwanda’s principal exports of coffee and tea, austerity measures imposed by the World Bank, and a drought in the south. Habyarimana took yet another attempted Tutsi invasion of northeastern Rwanda from neighboring Uganda in October 1990 as the pretext for rounding up or killing Hutu dissidents and Tutsi all over Rwanda, in order to strengthen his own faction’s hold on the country. The civil wars displaced a million Rwandans into settlement camps, from which desperate young men were easily recruited into militias. In 1993 a peace agreement signed at Arusha called for power-sharing and a multi-power government. Still, businessmen close to Habyarimana imported 581,000 machetes for distribution to Hutu for killing Tutsi, because machetes were cheaper than guns.

  However, Habyarimana’s actions against Tutsi, and his newfound toleration of killings of Tutsi, proved insufficient for Hutu extremists (i.e., Hutu even more extreme than Habyarimana), who feared having their power diluted as a result of the Arusha agreement. They began training their militias, importing weapons, and preparing to exterminate Tutsi. Rwandan Hutu fears of Tutsi grew out of the long history of Tutsi domination of Hutu, the various Tutsi-led invasions of Rwanda, and Tutsi mass killings of Hutu and murder of individual Hutu political leaders in neighboring Burundi. Those Hutu fears increased in 1993, when extremist Tutsi army officers in Burundi murdered Burundi’s Hutu president, provoking killings of Burundi Tutsi by Hutu, provoking in turn more extensive killings of Burundi Hutu by Tutsi.

  Matters came to a head on the evening of April 6, 1994, when the Rwandan presidential jet plane, carrying Rwanda’s President Habyarimana and also (as a last-minute passenger) Burundi’s new provisional president back from a meeting in Tanzania, was shot down by two missiles as it came in to land at the airport of Kigali, Rwanda’s capital, killing everyone on board. The missiles were fired from immediately outside the airport perimeter. It remains uncertain to this day by whom or why Habyarimana’s plane was shot down; several groups had alternative motives for killing him. Whoever were the perpetrators, Hutu extremists within an hour of the plane’s downing began carrying out plans evidently already prepared in detail to kill the Hutu prime minister and other moderate or at least less extreme members of the democratic opposition, and Tutsi. Once Hutu opposition had been eliminated, the extremists took over the government and radio and set out to exterminate Rwanda’s Tutsi, who still numbered about a million even after all the previous killings and escapes into exile.

  The lead in the killings was initially taken by Hutu army extremists, using guns. They soon turned to efficiently organizing Hutu civilians, distributing weapons, setting up roadblocks, killing Tutsi identified at the roadblocks, broadcasting radio appeals to every Hutu to kill every “cockroach” (as Tutsi were termed), urging Tutsi to gather supposedly for protection at safe places where they could then be killed, and tracking down surviving Tutsi. When international protests against the killings eventually began to surface, the government and radio changed the tone of their propaganda, from exhortations to kill cockroaches to urging Rwandans to practice self-defense and to protect themselves against Rwanda’s common enemies. Moderate Hutu government officials who tried to prevent killings were intimidated, bypassed, replaced, or killed. The largest massacres, each of hundreds or thousands of Tutsi at one site, took place when Tutsi took refuge in churches, schools, hospitals, government offices, or those other supposed safe places and were then surrounded and hacked or burned to death. The genocide involved large-scale Hutu civilian participation, though it is debated whether as many as one-third or just some lesser proportion of Hutu civilians joined in killing Tutsi. After the army’s initial killings with guns in each area, subsequent killings used low-tech means, mainly machetes or else clubs studded with nails. The killings involved much savagery, including chopping off arms and legs of intended victims, chopping breasts off women, throwing children down into wells, and widespread rape.

  While the killings were organized by the extremist Hutu government and largely carried out by Hutu civilians, institutions
and outsiders from whom one might have expected better behavior played an important permissive role. In particular, numerous leaders of Rwanda’s Catholic Church either failed to protect Tutsi or else actively assembled them and turned them over to killers. The United Nations already had a small peacekeeping force in Rwanda, which it proceeded to order to retreat; the French government sent a peacekeeping force, which sided with the genocidal Hutu government and against invading rebels; and the United States government declined to intervene. In explanation of these policies, the U.N., French government, and U.S. government all referred to “chaos,” “a confusing situation,” and “tribal conflict,” as if this were just one more tribal conflict of a type considered normal and acceptable in Africa, and ignoring evidence for the meticulous orchestration of the killings by the Rwandan government.

  Within six weeks, an estimated 800,000 Tutsi, representing about three-quarters of the Tutsi then remaining in Rwanda, or 11% of Rwanda’s total population, had been killed. A Tutsi-led rebel army termed the Rwandan Patriotic Front (RPF) began military operations against the government within a day of the start of the genocide. The genocide ended in each part of Rwanda only with the arrival of that RPF army, which declared complete victory on July 18, 1994. It is generally agreed that the RPF army was disciplined and did not enlist civilians to murder, but it did carry out reprisal killings on a much smaller scale than the genocide to which it was responding (estimated number of reprisal victims, “only” 25,000 to 60,000). The RPF set up a new government, emphasized national conciliation and unity, and urged Rwandans to think of themselves as Rwandans rather than as Hutu or Tutsi. About 135,000 Rwandans were eventually imprisoned on suspicion of being guilty of genocide, but few of the prisoners have been tried or convicted. After the RPF victory, about 2,000,000 people (mostly Hutu) fled into exile in neighboring countries (especially the Congo and Tanzania), while about 750,000 former exiles (mostly Tutsi) returned to Rwanda from neighboring countries to which they had fled (Plate 22).

  The usual accounts of the genocides in Rwanda and Burundi portray them as the result of pre-existing ethnic hatreds fanned by cynical politicians for their own ends. As summed up in the book Leave None to Tell the Story: Genocide in Rwanda, published by the organization Human Rights Watch, “this genocide was not an uncontrollable outburst of rage by a people consumed by ‘ancient tribal hatreds.’ . . . This genocide resulted from the deliberate choice of a modern elite to foster hatred and fear to keep itself in power. This small, privileged group first set the majority against the minority to counter a growing political opposition within Rwanda. Then, faced with RPF success on the battlefield and at the negotiating table, these few powerholders transformed the strategy of ethnic division into genocide. They believed that the extermination campaign would restore the solidarity of the Hutu under their leadership and help them win the war . . .” The evidence is overwhelming that this view is correct and accounts in large degree for Rwanda’s tragedy.

  But there is also evidence that other considerations contributed as well. Rwanda contained a third ethnic group, variously known as the Twa or pygmies, who numbered only 1% of the population, were at the bottom of the social scale and power structure, and did not constitute a threat to anybody—yet most of them, too, were massacred in the 1994 killings. The 1994 explosion was not just Hutu versus Tutsi, but the competing factions were in reality more complex: there were three rival factions composed predominantly or solely of Hutu, one of which may have been the one to trigger the explosion by killing the Hutu president from another faction; and the invading RPF army of exiles, though led by Tutsi, also contained Hutu. The distinction between Hutu and Tutsi is not nearly as sharp as often portrayed. The two groups speak the same language, attended the same churches and schools and bars, lived together in the same village under the same chiefs, and worked together in the same offices. Hutu and Tutsi intermarried, and (before Belgians introduced identity cards) sometimes switched their ethnic identity. While Hutu and Tutsi look different on the average, many individuals are impossible to assign to either of the two groups based on appearance. About one-quarter of all Rwandans have both Hutu and Tutsi among their great-grandparents. (In fact, there is some question whether the traditional account of the Hutu and Tutsi having different origins is correct, or whether instead the two groups just differentiated economically and socially within Rwanda and Burundi out of a common stock.) This intergradation gave rise to tens of thousands of personal tragedies during the 1994 killings, as Hutu tried to protect their Tutsi spouses, relatives, friends, colleagues, and patrons, or tried to buy off would-be killers of those loved ones with money. The two groups were so intertwined in Rwandan society that in 1994 doctors ended up killing their patients and vice versa, teachers killed their students and vice versa, and neighbors and office colleagues killed each other. Individual Hutu killed some Tutsi while protecting other Tutsi. We cannot avoid asking ourselves: how, under those circumstances, were so many Rwandans so readily manipulated by extremist leaders into killing each other with the utmost savagery?

  Especially puzzling, if one believes that there was nothing more to the genocide than Hutu-versus-Tutsi ethnic hatred fanned by politicians, are events in northwestern Rwanda. There, in a community where virtually everybody was Hutu and there was only a single Tutsi, mass killings still took place—of Hutu by other Hutu. While the proportional death toll there, estimated as “at least 5% of the population,” may have been somewhat lower than that overall in Rwanda (11%), it still takes some explaining why a Hutu community would kill at least 5% of its members in the absence of ethnic motives. Elsewhere in Rwanda, as the 1994 genocide proceeded and as the number of Tutsi declined, Hutu turned to attacking each other.

  All these facts illustrate why we need to search for other contributing factors in addition to ethnic hatred.

  To begin our search, let’s again consider Rwanda’s high population density that I mentioned previously. Rwanda (and Burundi) was already densely populated in the 19th century before European arrival, because of its twin advantages of moderate rainfall and an altitude too high for malaria and the tsetse fly. Rwanda’s population subsequently grew, albeit with ups and downs, at an average rate of over 3% per year, for essentially the same reasons as in neighboring Kenya and Tanzania (New World crops, public health, medicine, and stable political borders). By 1990, even after the killings and mass exilings of the previous decades, Rwanda’s average population density was 760 people per square mile, higher than that of the United Kingdom (610) and approaching that of Holland (950). But the United Kingdom and Holland have highly efficient mechanized agriculture, such that only a few percent of the population working as farmers can produce much of the food for everyone else, plus some surplus food for export. Rwandan agriculture is much less efficient and unmechanized; farmers depend on handheld hoes, picks, and machetes; and most people have to remain farmers, producing little or no surplus that could support others.

  As Rwanda’s population rose after independence, the country carried on with its traditional agricultural methods and failed to modernize, to introduce more productive crop varieties, to expand its agricultural exports, or to institute effective family planning. Instead, the growing population was accommodated just by clearing forests and draining marshes to gain new farmland, shortening fallow periods, and trying to extract two or three consecutive crops from a field within one year. When so many Tutsi fled or were killed in the 1960s and in 1973, the availability of their former lands for redistribution fanned the dream that each Hutu farmer could now, at last, have enough land to feed himself and his family comfortably. By 1985, all arable land outside of national parks was being cultivated. As both population and agricultural production increased, per-capita food production rose from 1966 to 1981 but then dropped back to the level where it had stood in the early 1960s. That, exactly, is the Malthusian dilemma: more food, but also more people, hence no improvement in food per person.

  Friends of mine who visited
Rwanda in 1984 sensed an ecological disaster in the making. The whole country looked like a garden and banana plantation. Steep hills were being farmed right up to their crests. Even the most elementary measures that could have minimized soil erosion, such as terracing, plowing along contours rather than straight up and down hills, and providing some fallow cover of vegetation rather than leaving fields bare between crops, were not being practiced. As a result, there was much soil erosion, and the rivers carried heavy loads of mud. One Rwandan wrote me, “Farmers can wake up in the morning and find that their entire field (or at least its topsoil and crops) has been washed away overnight, or that their neighbor’s field and rocks have now been washed down to cover their own field.” Forest clearance led to drying-up of streams, and more irregular rainfall. By the late 1980s famines began to reappear. In 1989 there were more severe food shortages resulting from a drought, brought on by a combination of regional or global climate change plus local effects of deforestation.

  The effect of all those environmental and population changes on an area of northwestern Rwanda (Kanama commune) inhabited just by Hutu was studied in detail by two Belgian economists, Catherine André and Jean-Philippe Platteau. André, who was Platteau’s student, lived there for a total of 16 months during two visits in 1988 and 1993, while the situation was deteriorating but before the genocide’s explosion. She interviewed members of most households in the area. For each household interviewed in each of those two years, she ascertained the number of people living in the household, the total area of land that it owned, and the amount of income that its members earned from jobs off the farm. She also tabulated sales or transfers of land, and disputes requiring mediation. After the genocide of 1994, she tracked down news of survivors and sought to detect any pattern to which particular Hutu ended up being killed by other Hutu. André and Platteau then processed this mass of data together to figure out what it all meant.