The data used in the GCRI are open-source and widely accepted by the scientific community.
They are provided by institutions such as the World Bank, the UN, and a variety of academic institutions such as universities as well as research projects.
This report presents the work done between February 2017 and September 2017, specifically focused on improving the documentation on the regression model. The present report describes on the one hand the regression model, including the input data and the model itself. On the other hand, it presents the statistical significance test and the matrix of confusion that have been performed, in order to get a highly detailed analysis of the performances of the model. The results of these analyses are presented in chapter 4 and 5. This report is part of a series of documentations produced in 2017 aiming at improving the GCRI models with greater transparency and robustness. This work contributes to enhancing the GCRI performance.
Document (pdf, 2MB)
Halkia, S., Ferri, S., Joubert-Boitat, I., Saporiti, F., Kaufmann, M. The Global Conflict Risk Index (GCRI): Regression model, data ingestion, processing and output methods, EUR 29046 EN, Publications Office of the European Union, Luxembourg, 2017, ISBN 978-92-79-77693-9, doi:10.2760/303651, JRC108767.
Previous versions of the GCRI have incrementally developed a statistical regression methodology for defining conflict. Based on literature review from the conflict science field, five theoretical risk areas were identified (see risk assessment factors above). Within these, a further distinction was made between concepts, which were then represented with individual variables. The variables used are all relatively stable, in that little change is to be expected from year to year. The risk-of-conflict is assessed by the composite model at country level based on these variables. It consists in computing raw data for creating each of the variables, and in compiling all of them into one final score (per country). Country profiles are finally produced, in which the composite model’s results are visualized, namely the final score of conflict risk and the background data used to calculate it using the variables. This report presents the work done between February 2017 and September 2017, which has focused on quality control of the dataset, and on improving the documentation of the composite model. While the work presented here shows great advances in reliability and reproducibility, there is still potential for improvements. The report is structured to present the three main aspects of the GCRI composite model: the data management, the variables and the model. While the first part gives an overview on the variables and the datasets used, each indicator used for all the variables is described in detail in the second part, including a general description, its relevance with regard to conflict risk, where the data is sourced and how the data is transformed. In the third part, the composite model is described and the results are presented through specific examples..
Document (pdf, 2MB)
Halkia, S., Ferri, S., Joubert-Boitat, I., Saporiti, F., Conflict Risk Indicators:Significance and Data Management in the GCRI, EUR 28860 EN, Publications Office of the European Union, Luxembourg, 2017, ISBN 978-92-79-75768-6, doi:10.2760/44005, JRC107996.
This technical report presents the methodology and code documentation for version 5 of the Global Conflict Risk Index. This release features changes to the data management system and imputation methods, and includes a new composite indicator. A reliable data management system is presented, which combines the data from various sources and imputes missing data. The reproduction process for the two step regression model of previous versions This document focuses on the technical aspects of producing the dataset and statistical output. For details regarding the theoretical framework, please see the previous scientific report, doi 10.2788/184.
Document (pdf, 2MB)
Smidt, M., L. Vernaccini, P. Hachemer, T. De Groeve; The Global Conflict Risk Index (GCRI): Manual for data management and product output; EUR 27908 EN; doi:10.2788/705817
This report describes the concept and methodology of the Global Conflict Risk Index (GCRI). The index was designed to display countries’ risk vis-à-vis violent conflicts, allowing for a comprehensive global scan to identify countries at risk of both internal and interstate disputes. To achieve this, it seeks to move beyond existing models and improve our understanding of risk assessment in three ways: (1) It offers a comprehensive definition of risk and risk of conflict, (2) it makes use of a new dataset on conflicts, and (3) it addresses three different dimensions of conflicts, enabling users to identify a country’s individual risk for conflicts over national power, on a subnational level, and in the international sphere. Twenty-two political, security, socio-economic and structural indicators underpin the calculation of conflict risk. Clustered in five pillars or risk areas, they take into account the most recent results of quantitative conflict research as well as expert input from conflict and country practitioners, and are combined in a statistical model to assess the likely conflict intensity over the next 4 years. The statistical analysis was supported by a methodology group of renowned conflict experts. The result is an estimate of countries at risk for highly violent conflicts in the near future, as well as an assessment of the intensity of existing conflicts based on changing structural metrics. These consider the contextual features and risks for each country with regard to its political cohesiveness, international integration, its socio-economic development level, geographical factors, and its security environment. The index provides an intuitive framework for the analysis of a country’s vulnerability and the characteristic features that favour conflict onset and escalation.
While the GCRI supports a proactive risk management framework and provides a sound basis for qualitative risk assessment, its data and results reflect the findings of scientific literature and a statistical model and are therefore not to be confused with in-depth analyses by country experts or policy recommendations.
De Groeve, T., P. Hachemer, L. Vernaccini; The Global Conflict Risk Index (GCRI): A Quantitative Model; EUR 36880 EN; doi:10.2788/184