Risk Assessment and Early Warning and Their Uses for Prevention: Three Mini Case Studies

Barbara Harff
Professor of Political Science Emerita,
U.S. Naval Academy

Three countries at risk of geno/politicide are sketched. The cases are identified using data on all states of the international community that have populations over 1 million. The basis for the analysis and the full risk list is in my article in Politorbis  (Harff 2009) and also posted on the GPANet website. First, though, I will refresh your memory on how we came to do systematic risk assessment and comment on the state of early warning. 

The idea of systematic early warning is at least 25 years old, the execution young

Of the dozen or so scholars worldwide who did comparative research on genocides in the early 1980s, six of us met  at the ISA meetings in St. Louis in 1988.  We knew then that we needed to spread the message that genocides were not unique (as most Holocaust scholars would have it in the 1980s). I had  just published my data set identifying 46 cases of geno/politicide  (Harff and Gurr 1988) which made it easier to persuade critics that the critical mass was achieved that would allow for systematic and not so systematic comparison of cases. This dataset consisted  of 46 mini case studies that identified types of victims (class, race, religion, ethnie)--the same for perpetrators, as well as essential information on timeframe in which the episodes occurred, number of casualties,  circumstances of death, and type of polity; it would eventually be part of the dataset of state failures compiled for the Political Instability Task Force (see http://globalpolicy.gmu.edu/pitf/ ).

 We understood that prevention is not possible if we do not understand the etiology of a given disease--to invoke a medical metaphor. We also needed to understand the causal chain that turns a high risk situation into a full-fledged genocide-the task for early warners.   

The idea of EW in cases of genocide was probably advanced first by Israel Charny, a clinical psychologist. In 1986 we had tried to combine forces in developing an early warning model that would encompass several levels of analysis.  His concerns were--true to his training--how group dynamics impact behavior and how individual motives, training, and psychological conditions could turn ordinary citizens into mass murderers.  Mine were similar but I was more interested in identifying local and national and international factors that enabled elites to commit crimes against humanity and get away with it.  Our efforts faltered. Why?  When we tried to account for all possible reasons why genocides occur the model became too convoluted to be tested and applied.  A more parsimonious model later developed in the late 1990s with the help of the Task Force yielded sufficiently strong and consistent results to permit solid risk assessment (Harff 2003).

The model is the result of testing relevant hypotheses derived from the conflict and genocide literature and focuses on structural factors.  We tested some 40 variables, ranging from economic indicators to political to environmental to demographic variables. The best-fit model identified the preconditions of over 80% of all post-1955 cases of geno/politicide.  The current risk assessments uses a somewhat better model fitted to the most recently available data. The methodology is explained in sources mentioned above-briefly what we asked was, why do some deadly and disruptive political conflicts-state failures-lead to genocides or politicides whereas most do not.

Just a very brief introduction to EW, it is different from risk assessment.  EW is supposed to tell us WHEN conflict is likely to turn into genocide.  In the late 1990s  I was asked by the US intelligence community to develop an EW model for geno/politicide.   I had developed  such a model  and applied it to data on events leading to mass atrocities in Bosnia, Rwanda, Burundi and Abkhazia (Harff 1996).

Mind you, I was not only fully aware of the concerns of area and country experts but shared at least some of their sentiments: structures cannot and do not tell the whole story.  However, I am skeptical about the importance of cultural factors. For example, the idea that something in the German culture led to blind obedience among Nazi followers does not really pan out.  Probably more persuasive is the fact that in a totalitarian police state individual heroism is costly for not just the individual but whole families and obedience or acceptance become the easy way out.  For political scientist regional, global factors matter more not to mention social organization, structures and authority patterns.  Culture is the fuzziest of all concepts and very hard to operationalize.

In the EW project: we tracked daily such variables as increasing militancy of rebels, increases and decreases in material, diplomatic and military aid to both rebels and regimes, hate propaganda,  and changes in state capacity. We also tracked increases and decreases in state discrimination and violence against specific groups. Moreover we were concerned with and monitored some international and regional state behaviors. Data were analyzed using standard and not so standard statistical methods.  How different was our effort from reports from the International Crisis Group?  Simply put, in our risk assessment, information was coded and interpreted using theoretical filters and then systematically  analyzed and reported.  And we were specifically concerned with genocides, politicides and mass atrocities, not generic conflict.  Much was learned from this effort.  For example, a key element that accelerates the genocidal process is increases in hate propaganda, as in Ivory Coast in the mid 2000's, and we know the outcome thanks to Juan Mendez's diplomatic efforts.

Now to the three countries at risk of future genocide and politicide.  The two highest risk cases are Sudan and Burma, familiar to all of you.  The third case is Saudi Arabia, identified by my risk analysis as at medium risk. First, Sudan and Burma were a long time in our sights.  Both have had historical geno/politicides--a key risk factor.  In Sudan during the civil wars victims were mostly Southerners; the campaigns against the Nuba and then Darfur were well known in advance of full-fledged assaults. The same goes for Arakanese Muslims in Burma.  Both cases rank highest on our structural risk factors.  (1) They are at high risk of future instability.[1] (2) Both practice state-led discrimination against one or more specific communal groups, as identified in the Minority at Risk dataset.  (3) Both are potentially repeat offenders, having committed other geno/politicides since 1950. Both have (4) ethnically polarized elites and (5) exclusionary ideologies. (These variables were derived from my original dataset on cases and re-coded for the Political Instability Task Force). (6) Both have autocratic regimes. And (7) trade openness is low in both countries, indicative of relatively low levels of international engagement.   

 There are other indicators of interest in these and other cases-indicators  derived from our EW research: They include increases in bribery and corruption, less spending on public goods affecting minorities,  militias that  increasingly take on the role of state authorities, and ideological encroachment on the judiciary.

In summary, both Burma and Sudan rank high on all relevant structural indicators, and medium on future instability-a very temporary situation. Burma's medium-low trade openness grade is due mostly to the large black market sector.  On EW indicators both rank high-questioning of course whether or not the rule of law means anything in both systems.

What do these rankings mean for prevention?  First, some general comments.

 Structural factors can be influenced by outsiders.  Take the variables of state-led discrimination, type of polity, and trade openness-the international community can have impact on seemingly intransigent regimes through pressure and incentives.  Take Southern Sudan's self-determination referendum, scheduled for January 2011. Is it possible to secure it? Given the lack of structures, the rainy season , lack of human security, and tension between the northern and southern governments, then international help is a must. Is it possible to preserve a relative peace between North and South in a united Sudan?  I think a clear separation is preferable, given the nature and length of past conflict.  However successful dissolution, without a new genocidal civil war, presupposes that the international community makes up its collective mind about who and what efforts to support.  Who should oversee or guarantee an orderly process for distributing oil revenues, or establishing contested borders?  If states are more so interested in preserving the status quo, i.e. the shaky peace between the North (the NCP) and the South (the SPLM), then massive help is needed to stop food shortages, provide for human security, and control rebel and militia activities.  Somehow the international community must find ways to break the cycle of mistrust between Northern and Southern leaders. And, as always, the UN and African Union must have the means and mandate to take sides in order to quell massive violence, if and when it resumes.

Burma is the most intransigent case in my opinion. There was never and may never be a unified Burma.  Whereas some regional minorities have made (at least temporary) peace with Rangoon, others remain at the fringes-discriminated, disenfranchised, and victims of ethnic repression.  I have argued for a long time that Burma needs to be fully included into the international community not excluded.  As mentioned above, internal change can only be accomplished if other states can gain entry or are at least are of some importance to intransigent regimes.  A refined carrots and sticks policy may work for China, Japan and ASEAN states and regional organizations that have a semblance of influence on Burmese politics. Neither the US nor Europe can have significant impact without prior coordination with our Asian partners. 

What are the risks now? The Rohingya Muslims are targets (again) of religious persecution and other human rights violation--leading to an increase of  Rohingyas seeking refuge in neighboring countries-despite restrictions on movement.  Discrimination against Chins  by local authorities is not just tolerated but actively encouraged by the ruling junta.  There is no independent judiciary.  The Shan states throughout history were either independent, aligned with Burmese monarchies, conquered by or aligned with China, or supported by Thailand (at least as far as non Communist Shan insurgent movements were concerned).  They continue their tortuous path to independence or at least formal autonomy.  Despite some overtures by Rangoon, not all Shan militias are likely to abide by the cease fire now in place nor will they disband.  The old Kuomintang bases are now home to many Chinese citizens doing brisk local business and extracting resources.  And the opium fields still support many local warlords.  

What to do?  Support and focus on ASEAN, Japan and China to work out some formula that would lessen oppression of ordinary peoples-there is no comprehensive solution for Burma.  And yes there is China-a country aggressively pursuing its international economic interests.  Interestingly enough its involvement in select African states tells a story.  They built highways, ports, factories, railways-in other words infrastructure where there was none.  In exchange they exploit and export needed natural resources.  Call it neo-colonialism-but it somehow works better in improving the lives of ordinary people.  They circumvent corrupt elites and avoid giving aid for projects that never get off the ground. They also build by bringing in Chinese labor to work with locals.  What can be learned from that model?   Get China involved in Burma via their economic interests-and recognize that aid without political conditions leads to apathy, corruption and dependence. 

Saudi Arabia as a medium risk case may come as a surprise to many.  Here we are not talking about the imminence of genocide but a possible (albeit not likely) politicide. So why bother talking about it?  This is a case that is not likely to make anybody else's list; but we  should worry about an increasingly radicalized Islam.  Wahabism is more than a puritan form of Islam;  it is the basis of a powerful political ideology.  Local radical clerics appear to gain influence and attract disgruntled youth vis a vis the moderates (and what, I ask you, do we here mean by moderates????)  Saudi royalty (including the king) seem to be no match for the growing influence of Wahabi/Hanbali doctrines in the region and  abroad-as far as Morocco in the west, Malaysia in the east, and in the Balkans in the north. We know that Saudi public and private foreign assistance has supported radical organizations that celebrate martyrdom aka terrorism.  Islamic militancy Wahabi style is also spreading throughout the Arabian peninsula.  In Yemen the Sa'ana leadership has little power over Northern tribal leaders nor the former South.  Wahabism has already increased illiteracy, suppression of women, and the control of education and information-eventually leading to more poverty, less development, and radicalization of tribes.   We need to build schools, de-salination plants, infrastructure, and give targeted aid to connect the backwaters of the Muslim world to the international community.  If the Saudis are our true friends-then they need to stop supporting radical causes and movements and control their own.  If  Wahabis ever come to power in Saudi Arabia or elsewhere then the likely victims of oppression or worse include women, foreigners, Christians, Jews, moderate and secular Arabs, Shi'a,  Alevis, Bahai's and still others.

In conclusion, let me be blunt. It would be great to have solid risk assessment and a reasonably well functioning EW system. These buy time for would-be intervenors or at least allow for better planning.  But here is the real problem for interested parties.  Any form of active involvement by international actors is driven by national interest and capacity.  Samantha Power calls it political will-but in my view this is a too narrow perspective. National interests are shaped by a combination of economic, political, and environmental/security/demographic factors as well as moral considerations. If national interest dictates some action then state capacity comes into play. And capacity means more than the possession of material means to act. States may be hindered or encouraged by history, law, and public pressure to engage.  And finally, it is necessary to assess realistically who our true friends are-neither Saudi Arabia or Pakistan should make the cut.



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

Harff, B. (1996) "Early Warning of Potential Genocide: The Cases of Rwanda, Burundi, Bosnia, and Abkhazia," ch. 3 in T. R. Gurr and B. Harff, Early Warning of Communal Conflicts and Genocide: Linking Empirical Research to International Responses.  Tokyo: United Nations University, Monograph Series on Governance and Conflict Resolution.

Harff, B., and T. R. Gurr (1988) "Toward Empirical Theory of Genocides and Politicides: Identification and Measurement of Cases since 1945," International Studies Quarterly, vol. 32, no. 3, pp. 359-371.

Hewitt, J. J. (2009) "The Peace and Conflict Ledger: Ranking States on Future Risks, 2008-2010." Ch. 3 in J. J. Hewitt, J. Wilkenfeld, and T. R. Gurr, Peace and Conflict 2010 (Boulder CO and London:  Paradigm Publishers).


[1] Future instability is rated using a composite index developed by J. J. Hewitt (2009) and described in T. R. Gurr's paper for the GPANet conference, ""A Legacy of Deadly Political Violence".


HarffonRAandPrevention.710.doc52 KB