Forecasting Social Unrest: A Machine Learning Approach
IMF Working Papers, November 5, 2021
Source details
- Canonical URL
- Forecasting Social Unrest: A Machine Learning Approach
Other formats
Bibliographic details
- Authors: Chris Redl, Sandile Hlatshwayo
- Published: November 5, 2021
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781557758873.001
Summary
- Produces a social unrest risk index for 125 countries covering a period of 1996 to 2020.
- Risk is defined as the probability of unrest in the following year derived from a machine learning model.
- The prediction model correctly forecasts unrest in the following year approximately two-thirds of the time.
- Shapley values indicate that key drivers of the predictions include high levels of unrest, food price inflation and mobile phone penetration.
Data and Model
- Coverage: 125 countries.
- Time span: 1996 to 2020.
- Predictors: over 340 indicators covering a wide range of macro-financial, socioeconomic, development and political variables.
- Outcome: probability of unrest in the following year as the basis for the social unrest risk index.
Model Performance and Validation
- Forecast accuracy: correctly forecasts unrest in the following year approximately two-thirds of the time.
- Explainability: uses Shapley values to identify key drivers consistent with previous findings in the literature.
Key Drivers Identified
- High levels of prior unrest.
- Food price inflation.
- Mobile phone penetration.
Content in this bundle
- Working Paper