Risk assessment and early warning systems

Post
Written By Ellen Stensrud

Several systems of risk assessment and early warning of mass atrocities exist, and atrocity crimes such as genocide and political mass violence rarely come as a complete surprise. This is an overview of the research field of risk assessment and early warning of genocide and political mass violence.

In identifying countries at major risk of such mass atrocities, we now have quite solid risk assessment systems. The vital task of early warning about whether and when mass atrocities will occur poses huge challenges, and there are several gaps to fill in the research field.

Risk assessment and early warning systems

Photo by Andrew Renneisen/Getty Images

Useful resources

Risk assessment and early warning: An overview

Since the end of the Cold War, think tanks have attempted to measure the risk of conflict and instability. The international failure to prevent mass atrocities in Rwanda and Bosnia spurred further initiatives of risk assessment, such as the US government’s Political Instability Task Force.[1] A groundbreaking contribution to the risk assessment of genocide and politicide[2] was Barbara Harff’s article from 2003, “No Lessons Learned from the Holocaust? Assessing Risks of Genocide and Political Mass Murder since 1955”.[3] Since then, many researchers and institutions have developed sets of indicators to predict genocide and similar episodes of mass violence.[4] Knowledge has informed advocacy work, often through mappings and lists of countries most at risk of atrocity crimes. One example is the Peoples under Threat (PUT) survey by Minority Rights Group International (MRG), another is the Early Warning Project.[5] A useful overview of the research field as well as policy options for atrocity prevention is presented by Scott Straus in his book Fundamentals of Genocide and Mass Atrocity Prevention.[6]

RA and EW: Important differences

The boundaries between risk assessment (RA) and early warning (EW) are not always clear-cut, but for analytical purposes a distinction can be useful. While RA indicators often focus on state-level characteristics such as regime type, past atrocities and war, EW indicators often point to more intermediate and short-term factors that may change rapidly, sometimes labelled “triggers”.[7] While RA usually relies on statistical and aggregated data or historical knowledge readily accessible for quantitative or comparative research, EW indicators are of a more dynamic and contextual nature, and local variations may be significant. For EW, country-specific expertise is therefore important. Moreover, RA models are predictive, not causal. They do not explain why atrocities take place, as the factors included in RA do not necessarily have a causal impact on the atrocities in question.[8] Since the end of the Cold War, think tanks have attempted to measure the risk of conflict and instability. The international failure to prevent mass atrocities in Rwanda and Bosnia spurred further initiatives of risk assessment, such as the US government’s Political Instability Task Force.[1] A groundbreaking contribution to the risk assessment of genocide and politicide[2] was Barbara Harff’s article from 2003, “No Lessons Learned from the Holocaust? Assessing Risks of Genocide and Political Mass Murder since 1955”.[3] Since then, many researchers and institutions have developed sets of indicators to predict genocide and similar episodes of mass violence.[4] Knowledge has informed advocacy work, often through mappings and lists of countries most at risk of atrocity crimes. One example is the Peoples under Threat (PUT) survey by Minority Rights Group International (MRG), another is the Early Warning Project.[5] A useful overview of the research field as well as policy options for atrocity prevention is presented by Scott Straus in his book Fundamentals of Genocide and Mass Atrocity Prevention.[6]

Risk assessment

  • Focuses on state level.

  • Looks at structures that do not change often.

  • Gives an overall picture of risk. Not when or if mass violence will erupt.

  • Predictive, not causal models.

Early warning

  • Monitoring of high-risk situations.

  • Escalation, tipping points and triggers.

  • More dynamic and contextual.

  • Intermediate/short-term.

In analyses of conflicts, RA and EW indicators can be linked to different phases of a conflict, although boundaries between such phases are blurred. RA indicators usually refer to structural features of a country that seldom change quickly. EW indicators often refer to developments or escalation once a conflict is underway in countries that are already on the RA radar and where the risk of mass atrocities is increasing. After identifying EW signals, these must be used to communicate to political actors in order to generate political will to address the crisis. In the research field, the three phases and challenges of risk assessment, early warning and generating the will to act are often treated separately, although in the real world, particularly for actors on the ground, the lines between the three may be difficult to identify.

Most RA lists are based on an assessment of all the countries in the world, although some, notably Harff (2003), are based on a selection of countries already in conflict. RA does not, however, say much about whether or when mass atrocities will occur. EW indicators can be used to look more closely at countries already considered at risk of mass atrocities to gain a fuller picture of the risk of atrocities and possibly to advocate for efforts to prevent them. Importantly, however, despite knowledge of EW indicators, it is often impossible to predict mass atrocities precisely. A systematic combination of RA and EW may still provide a good foundation for advocacy and warnings in order to generate political will to act.

Risk assessment and early warning systems
 

Risk assessment: Common and uncertain findings

In the scholarly field, there is general consensus on some of the RA indicators. These include instability and armed conflict; ideology of a transformative or exclusionary nature; and discrimination and unpunished violence against a potential victim group.[9] Together, these factors help single out cases where genocide or mass violence is most likely to occur.[10] Other findings are more disputed. These are used in some RA systems but not all, and must be regarded as uncertain.

Risk assessment indicators: Common findings

  • Major political instability (coup, revolution, civil war)

  • History of prior genocide or mass atrocities

  • Transformative/exclusionary ideology (revolutionary or nationalistic)

  • Past and ongoing state-led discrimination

One of the factors most strongly associated with the risk of mass atrocity or genocide is war or other types of political instability, such as revolution or coup. A situation of war creates a political climate of fear, uncertainty and legal and moral breakdown. Almost all cases of genocide or mass atrocity have occurred in the context of international or civil wars, or other instances of major instability. Another robust finding concerns prior mass violence against the victim group, although it is less clear why this is a factor that may predict genocide or mass atrocity. Past violence may simply say something about a political leadership’s willingness to resort to violence, or there may be a process of learning and escalation. The importance of ideology is linked to the decision making of political leaders. Revolutionary or nationalistic ideologies create a hierarchy of people and ideas about the necessity of radical social transformation that may motivate or justify genocide or mass atrocity. Finally, state-led discrimination places groups or individuals outside the “universe of moral obligation”, often seen as an important step towards genocide.

More uncertain RA indicators include inter-group hatred, a regime’s poor government capacity and authoritarian nature, and economic crises. Findings regarding these indicators are ambiguous and uncertain. Regarding regime type, recent cases indicate that transition periods entail greater risks of genocide and other mass violence than does stable authoritarianism.[11]

Potential and accuracy of RA systems

Some cautionary words regarding the accuracy of RA systems are called for. Despite scientific terminology such as “prediction” and “forecasting”, one should not underestimate the difficulties of predicting precisely when certain political events such as mass atrocities will occur. Further, these RA systems primarily refer to larger-scale episodes of mass violence, such as genocide, while smaller-scale instances of mass atrocities, such as the targeting of civilians in civil war, may be much harder to predict.[12] Moreover, most RA scholarship focuses on violence committed by state actors, but many atrocities are carried out by rebel or insurgent groups.[13] Finally, the complex and multifaceted processes that lead to mass atrocities can never fully be predicted, and researchers sometimes reach very different conclusions.[14] In scientific terms “false positives” will occur, since risk factors such as war, instability and repression are far more common than genocide and similar forms of large-scale violence. “False negatives” will also occur, but probably less often, since large-scale violence is the result of a long process, and usually occurs in countries generally regarded as unstable, repressive or at particular risk. Despite these weaknesses, the precision of some of the RA systems in predicting mass atrocities is still quite impressive. However, when it comes to determining when, how and why atrocities will take place, one must turn to more contextual and dynamic early warning indicators.

Early warning indicators

Regarding EW indicators, there is less agreement and certainty among researchers and fewer systematic efforts to evaluate their capability to predict mass atrocities. Many early warning models are developed by NGOs, international institutions or government agencies.[15] Further, EW indicators are, as explained above, more dynamic and country-specific than RA. EW indicators usually refer to short-term or intermediate changes and escalations. One can expect these to vary substantially between cases, according to the dynamics of each conflict. Therefore, these indicators are mostly based on single-case and comparative-case knowledge from scholars who have conducted research on specific genocides and episodes of mass atrocities. Early warning indicators can be divided into “triggers” and “escalation”, and may vary according to phase of conflict.[16]

Common early warning indicators

Leadership and regime

  • Polarisation of elites

  • Upcoming and contested elections

  • Public commemoration of past crimes

  • Rapid change in government leadership (assassination, coup)

  • Removing moderates from leadership

  • Attacks (arrests, torture, killings) on political leaders or other prominent figures

Discrimination and rhetoric

  • Increased hate speech, apocalyptic public rhetoric

  • Popular mobilisation against groups; labelling groups as enemies

  • Discriminatory or emergency legislation

  • Increase in repressive practices, removal of political rights

  • Segregation and separation of groups

Conflict dynamics

  • Increase in irregular armed forces and security forces, increase in opposition capacity

  • Increase in stockpiling and transfer of weapons

  • Commencement and resumption of armed conflict, spillover from neighbouring countries

  • Lack of opportunities to flee

  • Impunity for past crimes

Ways forward for early warning

As is evident from this list, there are too many EW indicators to create a predictive model. The relevant indicators will vary from case to case and in different phases of conflict. It is therefore not possible to argue that these indicators can be used to predict mass violence with a specific degree of certainty. Rather, they should be used as guidance on how to analyse countries that are already seen to be at high risk of mass violence. These EW indicators are potentially useful tools for civil society organizations and other actors who follow conflict situations closely, and who do advocacy work towards international bodies. Since the field of EW is less developed than that of RA, efforts should be directed at developing it further. One possible avenue for research in EW is to collect and compare data on periods just before or during an escalation. This could contribute to a more systematic and well-tested EW index.

This interview is produced by the project Inclusive Citizenship and Human Rights headed by Dr. Ingvill T. Plesner. www.inclusive-citizenship.no

References

"Early Warning Project." Accessed 16 January 2019. https://earlywarningproject.ushmm.org/.

Goldstone, Jack A., Robert H. Bates, David L. Epstein, Ted Robert Gurr, Michael B. Lustik, Monty G. Marshall, Jay Ulfelder, and Mark Woodward. "A Global Model for Forecasting Political Instability." American Journal of Political Science 54, no. 1 (2010): 190-208.

Harff, Barbara. "No Lessons Learned from the Holocaust? Assessing Risks of Genocide and Political Mass Murder since 1955." American Political Science Review 97, no. 1 (2003): 57-73.

Heldt, Birger. Mass Atrocities Early Warning Systems: Data Gathering, Data Verification and Other Challenges. In Genocide Prevention Advisory Network Conference. The Hague, 2012.

Straus, Scott. "Contested meanings and conflicting imperatives: A conceptual analysis of genocide." Journal of Genocide Research 3, no. 3 (2001/11/01 2001): 349-75.

Straus, Scott. Fundamentals of Genocide and Mass Atrocity Prevention.  Washington, D.C.: United States Holocaust Memorial Museum, 2016.

Verdeja, Ernesto. "Predicting Genocide and Mass Atrocities." Genocide Studies and Prevention: An International Journal 9, no. 3 (2016): 13-32.

 
 

[1] Ernesto Verdeja, "Predicting Genocide and Mass Atrocities," Genocide Studies and Prevention: An International Journal 9, no. 3 (2016): 15.
[2] This term is meant to cover mass violence directed against groups who are not protected under the Genocide Convention.
[3] Barbara Harff, "No Lessons Learned from the Holocaust? Assessing Risks of Genocide and Political Mass Murder since 1955," American Political Science Review 97, no. 1 (2003).
[4] This outline covers the RA and EW for the most severe cases of mass atrocities, such as genocide and political mass murder.
[5] "Early Warning Project," https://earlywarningproject.ushmm.org/.
[6] Scott Straus, Fundamentals of Genocide and Mass Atrocity Prevention (Washington, D.C.: United States Holocaust Memorial Museum, 2016).
[7] Ibid., 53-54.
[8] Verdeja, "Predicting Genocide and Mass Atrocities," 14-15.
[9] Straus, Fundamentals of Genocide and Mass Atrocity Prevention, 55-60. This list is a slightly revised version of Straus’ list. Goldstone et al. found regime characteristics to be by far the most important predictor. Jack A. Goldstone et al., "A Global Model for Forecasting Political Instability," American Journal of Political Science 54, no. 1 (2010): 194, 204.
[10] Straus, Fundamentals of Genocide and Mass Atrocity Prevention, 61.
[11] Ibid., 55-68.
[12] Heldt, Mass Atrocities Early Warning Systems: Data Gathering, Data Verification and Other Challenges, 13-14.
[13] Straus, Fundamentals of Genocide and Mass Atrocity Prevention, 68.
[14] Ibid., 54.
[15] Verdeja, "Predicting Genocide and Mass Atrocities," 24.
[16] This list is developed from Straus (2016) and Verdeja (2016).

Ellen Stensrud
Previous

Freedom of religion or belief (FoRB) and the Rohingya