Part 2 The Traps
Part 2 The Traps
CHAPTER 2 The Conflict Trap ALL SOCIETIES HAVE CONFLICT;it is inherent to politics.The prob- lem that is pretty distinctive to the bottom billion is not political conflict but its form.Some of them are stuck in a pattern of violent internal challenges to government.Sometimes the violence is prolonged,a civil war;sometimes it is all over swiftly,a coup d'etat.These two forms of political conflict both are costly and can be repetitive.They can trap a country in poverty. Civil War Seventy-three percent of people in the societies of the bottom billion have recently been through a civil war or are still in one.Many other countries have had civil wars at one time or another-the United States had one in the nineteenth century,Russia one early in the twentieth century,and Britain one back in the seventeenth-but,as these examples show,wars are not necessarily traps.The American,Russian,and British civil wars were ghastly at the time but were over fairly quickly and were not repeated.For low-income countries,however,the chances of war becoming a trap are much higher.I discovered this working with Anke Hoeffler,a young woman who was initially my doctoral student and is now my colleague. Anke's doctoral thesis was about the sources of growth,then a fashionable topic in economics.One of the factors known to impede growth is war.As I mulled over Anke's work it struck me that it would be interesting to turn
CHAPTER 2 The Conflict Trap All societies have conflict; it is inherent to politics. The problem that is pretty distinctive to the bottom billion is not political conflict but its form. Some of them are stuck in a pattern of violent internal challenges to government. Sometimes the violence is prolonged, a civil war; sometimes it is all over swiftly, a coup d’état. These two forms of political conflict both are costly and can be repetitive. They can trap a country in poverty. Civil War Seventy-three percent of people in the societies of the bottom billion have recently been through a civil war or are still in one. Many other countries have had civil wars at one time or another—the United States had one in the nineteenth century, Russia one early in the twentieth century, and Britain one back in the seventeenth—but, as these examples show, wars are not necessarily traps. The American, Russian, and British civil wars were ghastly at the time but were over fairly quickly and were not repeated. For low-income countries, however, the chances of war becoming a trap are much higher. I discovered this working with Anke Hoeffler, a young woman who was initially my doctoral student and is now my colleague. Anke’s doctoral thesis was about the sources of growth, then a fashionable topic in economics. One of the factors known to impede growth is war. As I mulled over Anke’s work it struck me that it would be interesting to turn
18 THE TRAPS the inquiry around:instead of explaining whether a country grew fast or slowly in terms of whether it was at war or peace,we could investigate whether proneness to war was explicable in terms of differences in growth. Causes of Civil War So what causes civil war?Rebel movements themselves justify their actions in terms of a catalogue of grievances:repression,exploitation,exclusion. Politically motivated academics have piled in with their own hobbyhorses, which usually cast rebels as heroes.I have come to distrust this discourse of grievance as self-serving.Sorting out the causes of civil war is difficult: historians cannot even agree on what caused the First World War.Most wars have multiple layers of causality:personalities,hatreds,mistakes.Our approach was to try to explain civil war statistically,looking at a range of possible causes:social,political,geographic,and economic. The first and most critical step in statistical research is getting satisfac- tory data.We found a comprehensive list of civil wars produced at the University of Michigan,for many years the world's leading center for data on such political questions.The Michigan definition of civil war is an internal conflict that involves at least 1,000 combat-related deaths,with each side incurring at least 5 percent of these deaths.(One advantage of using criteria devised by another researcher is that your results cannot be contaminated by the temptation to bend definitions so as to get the re- sults that you expect.)While the figure of 1,000 combat deaths is arbi- trary,the point of drawing a line is that there really is a big difference be- tween low-level communal violence in which,say,fifty people are killed and a war in which thousands get killed.We then matched this list of civil wars against a mass of socioeconomic data,country by country and year by year,with the goal of trying to determine the factors that affected the likelihood of a civil war developing in a given country within the next five years. Our work has proved controversial.In part this is because the people at- tracted to the academic study of conflict tend to be politically engaged and are sympathetic to the acute grievances enunciated by various rebel move- ments,who often adopt extreme measures to oppose governments that in- deed may be unsavory.To such academics,the whole idea of investigating
the inquiry around: instead of explaining whether a country grew fast or slowly in terms of whether it was at war or peace, we could investigate whether proneness to war was explicable in terms of differences in growth. Causes of Civil War So what causes civil war? Rebel movements themselves justify their actions in terms of a catalogue of grievances: repression, exploitation, exclusion. Politically motivated academics have piled in with their own hobbyhorses, which usually cast rebels as heroes. I have come to distrust this discourse of grievance as self-serving. Sorting out the causes of civil war is difficult: historians cannot even agree on what caused the First World War. Most wars have multiple layers of causality: personalities, hatreds, mistakes. Our approach was to try to explain civil war statistically, looking at a range of possible causes: social, political, geographic, and economic. The first and most critical step in statistical research is getting satisfactory data. We found a comprehensive list of civil wars produced at the University of Michigan, for many years the world’s leading center for data on such political questions. The Michigan definition of civil war is an internal conflict that involves at least 1,000 combat-related deaths, with each side incurring at least 5 percent of these deaths. (One advantage of using criteria devised by another researcher is that your results cannot be contaminated by the temptation to bend definitions so as to get the results that you expect.) While the figure of 1,000 combat deaths is arbitrary, the point of drawing a line is that there really is a big difference between low-level communal violence in which, say, fifty people are killed and a war in which thousands get killed. We then matched this list of civil wars against a mass of socioeconomic data, country by country and year by year, with the goal of trying to determine the factors that affected the likelihood of a civil war developing in a given country within the next five years. Our work has proved controversial. In part this is because the people attracted to the academic study of conflict tend to be politically engaged and are sympathetic to the acute grievances enunciated by various rebel movements, who often adopt extreme measures to oppose governments that indeed may be unsavory. To such academics, the whole idea of investigating 18 THE TRAPS
THE CONFLICT TRAP 19 statistically whether there is a relationship between objective measures of grievance and a propensity to rebel is taken to be more or less an insult, since they know there is one.Admittedly,we fanned the flames on occa- sion:we entitled one of our papers"Greed and Grievance"and another "Doing Well out of War,"implying that rebel motivations may just con- ceivably not be any more heroic than the governments they oppose.At the less politicized end of the academic profession,however,our work has been taken seriously and frequently cited.We reached the policy world- I was invited to address the General Assembly of the United Nations-and have been featured in the media. We were also asked to use our model to predict where the next civil wars would be-the CIA was apparently interested.But we were never that foolish.Our predictions might have been used as labels and thus likely to damage the very countries I was concerned to help;they might even have become self-fulfilling prophecies.More fundamentally,our model cannot be used for prediction.It can tell you what typically are the structural factors underlying proneness to civil war and-what is some- times more interesting-what seems not to be very important.From this, it can tell you the sort of countries that are most at risk.But it cannot tell you whether Sierra Leone will have another civil war next year.That depends upon a myriad of short-term events. The first link we found was between risk of war and initial level of in- come.Civil war is much more likely to break out in low-income coun- tries:halve the starting income of the country and you double the risk of civil war.One might ask whether we got the causality mixed up-is it just that war makes a country poor,rather than that poverty makes a country prone to war?In fact,both relationships hold simultaneously.While civil war reduces income,low income indeed heightens the risk of civil war. The clearest evidence for this arises because during colonialism many countries experienced decades of enforced peace;the near-simultaneous decolonization of many countries with very different income levels pro- vided a natural experiment for the effect of income on civil war. The relationship between low income and civil war may seem obvious- if you read the newspapers,you will see that the countries where there is conflict are far more likely to be poor-but not all theorists of civil war have based their work on empirical data.Some social scientists,particularly the
statistically whether there is a relationship between objective measures of grievance and a propensity to rebel is taken to be more or less an insult, since they know there is one. Admittedly, we fanned the flames on occasion: we entitled one of our papers “Greed and Grievance” and another “Doing Well out of War,” implying that rebel motivations may just conceivably not be any more heroic than the governments they oppose. At the less politicized end of the academic profession, however, our work has been taken seriously and frequently cited. We reached the policy world— I was invited to address the General Assembly of the United Nations—and have been featured in the media. We were also asked to use our model to predict where the next civil wars would be—the CIA was apparently interested. But we were never that foolish. Our predictions might have been used as labels and thus likely to damage the very countries I was concerned to help; they might even have become self-fulfilling prophecies. More fundamentally, our model cannot be used for prediction. It can tell you what typically are the structural factors underlying proneness to civil war and—what is sometimes more interesting—what seems not to be very important. From this, it can tell you the sort of countries that are most at risk. But it cannot tell you whether Sierra Leone will have another civil war next year. That depends upon a myriad of short-term events. The first link we found was between risk of war and initial level of income. Civil war is much more likely to break out in low-income countries: halve the starting income of the country and you double the risk of civil war. One might ask whether we got the causality mixed up—is it just that war makes a country poor, rather than that poverty makes a country prone to war? In fact, both relationships hold simultaneously. While civil war reduces income, low income indeed heightens the risk of civil war. The clearest evidence for this arises because during colonialism many countries experienced decades of enforced peace; the near-simultaneous decolonization of many countries with very different income levels provided a natural experiment for the effect of income on civil war. The relationship between low income and civil war may seem obvious— if you read the newspapers, you will see that the countries where there is conflict are far more likely to be poor—but not all theorists of civil war have based their work on empirical data. Some social scientists, particularly the THE CONFLICT TRAP 19
20 THE TRAPS most politically engaged,know what they want to see in civil war and duly see it. What else makes a country prone to civil war?Well,slow growth,or worse,stagnation or decline.As an approximation,a typical low-income country faces a risk of civil war of about 14 percent in any five-year pe- riod.Each percentage point added to the growth rate knocks off a per- centage point from this risk.So if a country grows at 3 percent,the risk is cut from 14 percent to 11 percent;if its economy declines at 3 percent, the risk increases to 16 percent.On this point too,one might ask whether we have the causality backward-might it be the case instead that it is the anticipation of civil war that causes decline?After all,when a civil war looks to be in the cards,investors flee,and the economy declines.It looks like decline causes war,but actually it's the anticipation of war that causes decline.This objection can be dealt with by looking at a factor that affects growth but has no direct connection to civil war,and seeing whether the subsequent effects make civil war more or less likely.In low-income coun- tries rainfall shocks(too much or too little rain)affect economic growth, but they do not directly affect the risk of civil war-that is,prospective rebels do not say,"It's raining,let's call off the rebellion."The effects on growth of rainfall shocks are thus clean of any ambiguity:they are not caused by anticipation of civil war.Yet setbacks to growth caused by rain- fall shocks make civil war much more likely. So if low income and slow growth make a country prone to civil war, it is reasonable to want to know why.There could be many explanations. My guess is that it is at least in part because low income means poverty, and low growth means hopelessness.Young men,who are the recruits for rebel armies,come pretty cheap in an environment of hopeless poverty. Life itself is cheap,and joining a rebel movement gives these young men a small chance of riches.In 2002 a little gang of rebels in the Philippines managed to kidnap some foreign tourists.A French woman among the kidnapped later described how she wrote down their demands for trans- mission to the authorities."What do you want me to write?"she asked. "A million dollars per tourist"was what they wanted.She wrote it down, then asked,"Anything else?"A long pause,then a political thought: "Sack the mayor of Jolo."The last demand:"Two divers'wristwatches." That was the list of "totally justified"grievances from that particular
most politically engaged, know what they want to see in civil war and duly see it. What else makes a country prone to civil war? Well, slow growth, or worse, stagnation or decline. As an approximation, a typical low-income country faces a risk of civil war of about 14 percent in any five-year period. Each percentage point added to the growth rate knocks off a percentage point from this risk. So if a country grows at 3 percent, the risk is cut from 14 percent to 11 percent; if its economy declines at 3 percent, the risk increases to 16 percent. On this point too, one might ask whether we have the causality backward—might it be the case instead that it is the anticipation of civil war that causes decline? After all, when a civil war looks to be in the cards, investors flee, and the economy declines. It looks like decline causes war, but actually it’s the anticipation of war that causes decline. This objection can be dealt with by looking at a factor that affects growth but has no direct connection to civil war, and seeing whether the subsequent effects make civil war more or less likely. In low-income countries rainfall shocks (too much or too little rain) affect economic growth, but they do not directly affect the risk of civil war—that is, prospective rebels do not say, “It’s raining, let’s call off the rebellion.” The effects on growth of rainfall shocks are thus clean of any ambiguity: they are not caused by anticipation of civil war. Yet setbacks to growth caused by rainfall shocks make civil war much more likely. So if low income and slow growth make a country prone to civil war, it is reasonable to want to know why. There could be many explanations. My guess is that it is at least in part because low income means poverty, and low growth means hopelessness. Young men, who are the recruits for rebel armies, come pretty cheap in an environment of hopeless poverty. Life itself is cheap, and joining a rebel movement gives these young men a small chance of riches. In 2002 a little gang of rebels in the Philippines managed to kidnap some foreign tourists. A French woman among the kidnapped later described how she wrote down their demands for transmission to the authorities. “What do you want me to write?” she asked. “A million dollars per tourist” was what they wanted. She wrote it down, then asked, “Anything else?” A long pause, then a political thought: “Sack the mayor of Jolo.” The last demand: “Two divers’ wristwatches.” That was the list of “totally justified” grievances from that particular 20 THE TRAPS