Bayesian spatial risk mapping of subnational conflict in Nigeria

This study asks how economic and geospatial factors shape different types of conflict within Nigeria and presents a high-resolution risk methodology that combines spatial econometrics with the greed–grievance framework. Using a Bayesian spatial approach, the authors model how socio-economic development, natural resources, energy infrastructure and multiple spatial predictors relate to four conflict typologies over 1997–2023, producing predictive risk maps intended to inform planning and stabilization.

Bayesian spatial model, Subnational high-resolution data, Four conflict typologies, Predictive risk maps

Data and methods

The authors analyze conflict in Nigeria between 1997 and 2023 and report a Bayesian approach that uses integrated nested Laplace approximation and stochastic partial differential equations to implement spatial econometric models. The design integrates the greed–grievance theoretical frame with 14 spatial variables including measures of wealth, ethnic fractionalization, resource endowments and power infrastructure.

Main findings

The abstract reports that wealth is associated with reduced conflict risk while ethnic fractionalization and proximity to resources show actor-specific effects. Petroleum endowments and power infrastructure are linked with higher organized rebel and militia activity, and ethnic dynamics are said to primarily drive riots. The authors also produce predictive risk maps to support infrastructure planning, supply-chain risk mitigation and targeted stabilization policies.

• Wealth reduces conflict risk (aggregate association reported).
• Ethnic fractionalization and proximity to resources have actor-specific effects.
• Petroleum endowments and power infrastructure associate with more organized rebel/militia activity; ethnic dynamics mainly drive riots.