What is the best way to assess the risk of state failure? What are the key indicators that a state is likely to fail? Authored by a member of the United States Government’s Political Instability Task Force (PITF), this Council on Foreign Relations paper draws heavily on PITF’s research and modelling. PITF’s recent models are 80 to 90 per cent accurate in predicting state failure.
PITF aims to predict the onset of political instability in a state two years in advance. It defines political instability as the occurrence of one or more political instability events. These events fall into four categories: ethnic war, revolutionary war, genocide or politicide, and adverse regime change. PITF does not attempt to predict which type of event will occur, rather whether one or more of these events are likely to occur in a particular state.
The four key indicators identified by the PITF global model as predictive of political instability are: regime type, infant mortality, armed conflict in neighbouring countries and state-led political discrimination. The most important indicator is regime type, specifically the condition of factionalism. In other words, political instability is preceded by periods of contentious politics.
- PITF’s models differ from many academic risk assessment models in that they assume probabilistic, rather than causal, relationships between indicators and outcomes.
- Indicators are selected based on the strength of their contribution to the predictive accuracy of the model, meaning that fewer indicators are included.
- PITF originally expected that no simple model would succeed in identifying indicators associated with the onset of political instability. But in fact there was significant agreement globally.
The identification of factionalism as a key indicator of political instability creates a window of opportunity for policymakers. If international actors intervene when factionalism occurs, it should be possible to prevent the onset of ethnic war, genocide or regime change. However, while there is significant agreement across PITF’s models, there are some differences across regions and types of instability:
- The PITF sub-Saharan Africa model shares with the global model the key indicators of regime type and state-led discrimination. However, other key indicators that political instability will occur include colonial history and leader’s tenure in office.
- The PITF Muslim countries model shares with the global model the key indicators of regime type, neighbouring countries with armed conflicts, and infant mortality.
- However, the PITF Muslim countries model does not have factionalism as a key indicator of political instability due to the relative absence of open, democratic regimes.
- The PITF ethnic war model shares many of the global model’s key indicators for political instability, however it includes a key indicator of conflict recurrence and of system manageability (population size and ethnic diversity).
