Civil war

• 1. Introduction

  ›

  Civil war haunts the modern world in a way that few phenomena can match. Since 1945, conflicts fought within a single country's borders have killed more than 25 million people and forced millions more from their homes. In the 1990s alone, roughly twenty civil wars were being waged simultaneously in any given year. That is ten times the historical average recorded since the 19th century.

  The word "civil" itself comes from the Latin for citizen. A civil war is a war between the citizenry, not against an outside enemy. Yet that distinction, which sounds simple, turns out to be extraordinarily difficult to pin down in law, in data, and in human experience.

  Who counts as a combatant? How many deaths cross the threshold from civil unrest into civil war? Does outside help from a foreign power change what a conflict is? And why do civil wars, once started, now last so much longer than they used to? Those questions sit at the center of one of the most contested fields in political science, and the answers carry real consequences for the people living through them.

• 2. The Problem of Definition

  ›

  James Fearon, a scholar of civil wars at Stanford University, defines the term as a violent conflict within a country fought by organized groups that aim to take power at the center or in a region, or to change government policies. Stathis Kalyvas narrows the focus further, specifying that the fighting must take place within a recognized sovereign entity between parties that were subject to a common authority when hostilities began.

  Ann Hironaka adds one more condition: at least one of those parties must be the state itself.

  Where scholars disagree most sharply is on the question of scale. Some political scientists draw the line at more than 1,000 total casualties. Others require at least 100 dead on each side, not just in the aggregate. The Correlates of War dataset, widely used by researchers, sets the bar at over 1,000 war-related casualties per year of conflict, not per conflict in total. That distinction matters enormously. Applied consistently, the per-year threshold excludes conflicts like the Troubles in Northern Ireland and the struggle of the African National Congress against apartheid in South Africa, yet it captures only a small fraction of the millions who died in the Second Sudanese Civil War or the Cambodian Civil War.

  Using the 1,000-casualties-per-year standard, researchers have counted 213 civil wars between 1816 and 1997, with 104 of those occurring in the compressed period from 1944 to 1997. Switching to the less demanding 1,000 total casualties threshold, over 90 civil wars took place between 1945 and 2007, with 20 still ongoing as of that year.

  The Geneva Conventions sidestep a formal definition entirely. Their text addresses "armed conflict not of an international character" without specifying what a civil war is. The International Committee of the Red Cross acknowledged this ambiguity directly, noting in its commentary that the relevant language was so general that many delegations feared it could be applied to almost any armed act.

• 3. Greed, Grievance, and Opportunity

  ›

  A team from the World Bank undertook one of the most comprehensive attempts to identify the causes of civil war in the early 21st century. Their framework, which became known as the Collier-Hoeffler Model, examined 78 five-year periods in which civil war occurred from 1960 to 1999, comparing them against 1,167 five-year periods of no civil war, then ran regression analysis to isolate which factors actually predicted conflict.

  The single strongest predictor was economic, not ethnic. A country at what the authors called "peak danger", where primary commodity exports made up 32% of gross domestic product, faced a 22% chance of falling into civil war in any given five-year window. A country with no primary commodity exports faced just a 1% risk. The explanation the researchers offered was bluntly practical: gold mines and oil fields are far easier to seize and control than garment factories or hotels.

  Diasporas matter too. The model found that switching a country's diaspora from the smallest in the study to the largest produced a sixfold increase in the probability of civil war, because diaspora communities can fund rebellions from abroad long after domestic resources run dry.

  Education and economic growth worked in the opposite direction. A male secondary school enrollment rate 10 percentage points above the study average reduced conflict probability by around 3%. A growth rate 1 percentage point higher than average reduced it by roughly 1%. The researchers read those numbers as reflecting opportunity cost: young men, who make up the vast majority of combatants, are less likely to pick up a weapon if they have an education, a salary, and a reasonable expectation of a stable future.

  The study's findings on identity-based grievances were more surprising. Ethnic polarization, religious fractionalization, and economic inequality were statistically insignificant predictors. Only ethnic dominance, meaning a country in which the largest ethnic group makes up a majority, nearly doubled the risk of civil war. Ethnic and religious diversity by itself did not increase risk; it slightly reduced it, as long as no single group dominated.

  David Keen, a professor at the Development Studies Institute at the London School of Economics, pushed back hard on what he saw as the model's reductiveness. He argued that conflicts are too complex to be explained by a single motive and that the quantitative methods used by Paul Collier miss the human texture of what drives people to fight. Cristina Bodea and Ibrahim Elbadawi went further, co-writing a piece titled "Riots, Coups and Civil War: Revisiting the Greed and Grievance Debate", arguing that empirical data can make greed theory largely irrelevant.

  James Fearon and David Laitin, whose 2003 article in the American Political Science Review is considered the most influential opportunity-based explanation for civil war onset, found that ethnic and religious diversity on its own does not make civil war more likely. What does make conflict more likely, they argued, is anything that helps rebels recruit and sustain themselves: poverty, politically unstable governments, rough terrain, and large populations.

• 4. How Long Wars Last

  ›

  Civil wars between 1900 and 1944 lasted an average of one and a half years. Since 1945, the average has risen to just over four years. That shift is not simply a change in severity; it has compounded over time, producing a world where many more wars are active simultaneously at any given moment.

  Ann Hironaka, in her book Neverending Wars, traces this arc across three historical periods. In 19th-century Europe, civil wars were typically short because they resolved around a clear center of power: whoever held the capital and commanded the military could crush resistance quickly. The 1871 Paris Commune is the illustration she uses. The fighting unfolded almost entirely within Paris and ended once the military sided with the government at Versailles and retook the city.

  The two major exceptions in the earlier period reveal the conditions that allow wars to lengthen. The American Civil War of 1861-1865 was unusual because it was fought over regional identities alongside political ideologies, and it ended not through a decisive battle for the capital but through attrition. The Spanish Civil War of 1936-1939 dragged on because both sides received outside support: Germany, Italy, and Portugal backed Francisco Franco, while France and the Soviet Union backed the government.

  After 1945, those exceptional conditions became the norm. Outside intervention became routine, and it had a measurable effect on duration. On average, a civil war with international intervention lasted 300% longer than one without. When intervention occurred on only one side, wars lasted 156% longer. When both sides received support, wars were longer by an additional 92% on top of that. If one of the intervening powers was a superpower, add another 72%. The Angolan Civil War, which saw two-sided foreign involvement including by two superpowers, illustrates the outer limit: a conflict of that configuration would be 538% longer on average than an unintervened war.

  The Cold War made matters structurally worse. The conflict between 1947 and 1991 created a global network of ideological and material support that kept wars alive in weak ex-colonial states. Civil wars that included pro- or anti-communist forces lasted 141% longer than non-Cold War conflicts. Wars that attracted actual superpower intervention lasted typically more than three times as long. When the Cold War ended, marked by the fall of the Berlin Wall in 1989, the duration of Cold War-associated civil wars fell by 92%, roughly a tenfold increase in the rate of resolution. Long-running conflicts in Guatemala, El Salvador, and Nicaragua reached their conclusions in short succession.

• 5. Weak States and the Colonial Legacy

  ›

  After World War II, the number of independent states jumped from roughly 30 ex-colonial nations to almost 120. Hironaka's statistical analysis estimates that the proliferation of ex-colonial states increased the post-1945 incidence of civil wars by 165% over the pre-1945 baseline.

  The reason comes down to a structural mismatch. Western states had spent centuries building administrative capacity through intense predatory competition with one another. The historian Charles Tilly captured this in a phrase that has since become widely cited: "War made the state and the state made war." The formation of Germany and Italy as modern states in the 19th century, for example, is directly associated with the wars of expansion led by Prussia and Sardinia-Piedmont respectively.

  The ex-colonial states received the outward forms of statehood: a flag, an anthem, a seat at the United Nations, a defined border, an official economic policy. What they did not receive was the bureaucratic infrastructure to make any of it work. Belgium prohibited any self-government in its colonies until it abruptly granted independence in 1960. Britain treated British Somaliland as little more than a trading post. France made all major decisions for its colonies in Paris. Most colonies were economic loss-makers at the moment of independence, without a productive economic base or a functioning tax system.

  In that environment, factions could manipulate the state apparatus to benefit themselves, and state leaders could use bureaucracies to serve personal interests rather than public ones. The newly independent states thus found themselves expected to perform all the functions of a modern state from day one, with almost none of the institutional capacity to do so.

  The international community then made an unintended choice that compounded the problem. Instead of allowing weak states to fragment or be absorbed, post-1945 international norms extended de jure sovereignty to every recognized state regardless of its de facto capacity. Weak states received diplomatic recognition, development aid, and protection from annexation. Anti-secessionist norms meant that rebel groups seeking to break away faced international opposition even when they held territory. Hironaka's study found that each major international anti-secessionist declaration was correlated with a 10% increase in the number of ongoing civil wars, adding up to a 114% increase from 1945 to 1997.

• 6. What Stops Wars From Ending

  ›

  Barbara F. Walter, a political scientist whose work examines why civil wars recur, argues that most contemporary civil wars are repeats of earlier ones. The conditions she identifies as drivers are not democracy or poverty per se, but the absence of accountable leadership, low public participation in politics, and a lack of information transparency between executives and the public. When those conditions are reversed, she argues, they act as legal and political constraints on executive power, create legitimate channels for political participation, and strengthen the credibility of peace commitments. The strength of a nation's institutions, in her view, is the primary indicator of whether a civil war will repeat.

  Commitment problems play a structural role in this dynamic. When two sides negotiate a peace agreement, both are aware that taking steps to demobilize or weaken their military position creates incentives for the other side to exploit the moment. Neither can credibly commit to holding its end of the bargain once the balance of power shifts. States with weak political and legal institutions lack the mechanisms to make such commitments credible or to enforce peace settlements once they are reached.

  Population geography and terrain also shape how conflicts develop once they begin. High population dispersion toward a country's borders makes it harder for a government to maintain control, since a dispersed population is inherently more difficult to monitor and police than one concentrated in a central region. Mountainous terrain provides rebels with sanctuaries. A 2006 systematic review identified rough terrain as one of the more important structural factors in civil war dynamics.

  Time itself has a pacifying effect, according to research on recurrence. The longer the period since a country's last civil war, the less likely a new one becomes. Two explanations have been proposed: either the material capital that made rebellion viable gradually loses value, or old hatreds slowly heal. The presence of a diaspora, however, substantially reduces the positive effect of elapsed time, because diaspora funding can offset the depreciation of whatever made rebellion viable in the first place.

  The economic costs compound across borders. Two separate studies estimate that each year of civil war reduces a country's GDP growth by roughly 2%. The damage does not stop at the border: civil wars reduce the GDP growth of neighboring countries as well, creating regional economic drag that can outlast the conflict itself.