Per capita income (logged, lagged) is one of the strongest predictors of civil war onset: a one standard deviation increase reduces onset probability by roughly half.
// strong — one SD increase (~1.5 log units) ≈ 50% reduction in onset odds
Fearon & Laitin (2003) "Ethnicity, Insurgency, and Civil War" — foundational paper on civil war onset. Challenges ethnic grievance explanations, finds insurgency-opportunity factors (poverty, terrain, weak states) are stronger predictors.
Per capita income (logged, lagged) is one of the strongest predictors of civil war onset: a one standard deviation increase reduces onset probability by roughly half.
Ethnic fractionalization (ELF index) is NOT a significant predictor of civil war onset once per capita income, terrain, and state capacity are controlled for.
Mountainous terrain significantly increases civil war onset probability. Log mountainous terrain has a positive and significant coefficient. Mountains provide insurgents with sanctuary and raise the costs of counterinsurgency.
Domain: Intra-State Conflict
Study of civil war onset, duration, and termination within sovereign states. Examines structural, economic, political, and geographic determinants of armed conflict.
Temporal scope: 1945-present | Population: Sovereign states (country-year observations)
…and 3 more findings
Per capita income (logged, lagged) is one of the strongest predictors of civil war onset: a one standard deviation increase reduces onset probability by roughly half.
Ethnic fractionalization (ELF index) is NOT a significant predictor of civil war onset once per capita income, terrain, and state capacity are controlled for.
Mountainous terrain significantly increases civil war onset probability. Log mountainous terrain has a positive and significant coefficient. Mountains provide insurgents with sanctuary and raise the costs of counterinsurgency.
Population size (logged, lagged) increases civil war onset risk. Larger countries have more potential recruits and greater heterogeneity.
Political instability (3+ point Polity change in prior 3 years) significantly increases civil war onset. Regime transitions create windows of vulnerability.
Anocracy (mixed regime, Polity -5 to +5) is positively associated with civil war onset, consistent with the inverted-U hypothesis. Neither full democracies nor full autocracies have elevated onset risk compared to anocracies.
Oil exporters have higher civil war onset probability. Oil rents reduce the need for broad-based taxation, weaken state-society ties, and may create prize-capture incentives for rebel groups.
New states (independent less than 2 years) have substantially elevated civil war risk. State formation and decolonization create windows of institutional fragility.
Noncontiguous territory increases civil war onset. Separated territories are harder to control militarily and may harbor aggrieved groups beyond state reach.
Religious fractionalization is not a significant predictor of civil war onset when controlled for other factors. Religion per se does not drive conflict risk.
The key mechanism for civil war is insurgency feasibility, not ethnic grievance. Conditions that make guerrilla warfare viable (rough terrain, weak states, poverty for cheap recruits) matter more than ethnic diversity or polarization.
name: fearon-laitin-2003 version: 1.0.1 pax_type: paper published_by: Praxis Agent domain: intra_state_conflict constructs: - mountainous_terrain - political_instability - noncontiguous_territory - oil_exporter - new_state - religious_fractionalization - per_capita_income - population_size - anocracy - civil_war_onset - ethnic_fractionalization engines: counts: constructs: 11 findings: 11 propositions: 0 playbooks: 2 sources: 1