During the Vietnam War, American soldiers destroyed the village of Ben Tre killing over five hundred men, women, and children in an effort to kill guerrillas (rebel soldiers) believed to be hiding there – three weapons were found (Taslitz, 2011). An American Major later justified the massacre by saying, “it became necessary to destroy the village in order to save it” (Taslitz, 2011).

War is often regarded as an idealized version of interstate conflict, with clear objectives, clear targets, and a clear enemy. The version of war so often pictured in the collective conscience reflects conflicts of the past more closely, where more often than not there were two distinct armies fighting tirelessly for a given outcome. However, the reality of the ‘modern war’ is much more chaotic.

The number of armed conflicts has increased since 2013 (Peace Research Institute Oslo, 2016). With this increase, there comes a distinct set of issues closely associated with this type of violence. Guerrilla warfare especially, is a rising means to combat a government in power inside of a civil conflict. Often, the technique is resorted to as a way to compensate for the asymmetric balance of power between citizen based rebel groups and a government regime.

Studying the already muddled arena of civil conflict is therefore made more hectic by the addition of these malleable, moveable rebel groups. The chaotic nature of civil conflict is transferred into its study, enforced by the lack of a comprehensive data on the location of these rebel groups. Knowing where rebel groups are located and how they move is important to study the patterns of civil conflict. For example, if rebel groups are consistently perpetrating violence in specific areas of their territory, defenses can be taken to counteract this. Consequently, in order to conduct any sort of spatial study on rebel groups, a bit of ingenuity is necessary.

Building on the work done by PhD candidates Philip Martin and Nina McMurry in the MIT Political Science Department, I created a dataset this summer to help solve this problem. Their work investigates the conditions under which humanitarian aid delivered after a ceasefire is most likely to hasten conflict resumption.

Martin and McMurry argue that previous assertions are correct in finding a destabilizing effect of humanitarian aid, but contest the conditions under which the proposition is correct. They find that increased levels of humanitarian aid to countries with civil war ceasefires is associated with a significant increase in the likelihood of ceasefire failure. In this case, humanitarian aid is defined “as food and medical supplies, as well as other support services and forms of assistance ‘designed to save lives, alleviate suffering and maintain and protect human dignity…” (Martin & McMurry, 2016, 14). Concluding that the destabilizing effect is catalyzed when rebel organizations controlled territory during the conflict, and proposing that this is due to territorially-based rebels’ greater ability to capture and leverage aid to continue their struggle.

My work will allow further testing of their argument. For example, if rebels’ access to territorial “safe havens” exacerbates the destabilizing effects of humanitarian aid, we should see those effects increase when rebel-held territory is closer to an international border or farther from the capital. Incorporating spatial data that gives us an idea of where rebels are located makes it possible to more rigorously evaluate the authors’ argument.

I began constructing a dataset that resembles rebel held territory by using ethnic association of rebel groups as a proxy for the location of their territorial control. This technique involved merging four datasets together (Geo-referencing of Ethnic Groups, Fortuna Peacekeeping and the Peacekept, UCDP/PRIO, ACD2EPR) to find the rebel groups associated with ceasefires, the ethnicities associated with those rebel groups, and where those ethnicities are located spatially to plot polygons of these ethnic groups.

Using GIS software, the data assembled is employed to construct shape files for each ethnically based rebel group and calculate the distance from those polygons to the respective border and capital of each associated nation.

Fifteen different ceasefires were randomly chosen to delve further in the literature associated with each in order to find the main provinces or areas controlled by the groups and if these areas match the polygons created. This spot checking procedure is used as a means to test if ethnicity can be used as a reliable proxy for rebel held territory.

For example, three main rebel groups were enlisted in the Ethiopian ceasefire of 1991 when Ethiopian Peoples’ Revolutionary Democratic Front (EPRDF) rebels took the capital and their allies Eritrean Peoples’ Liberation Front (EPLF) took control of Eritrea. Here each of the three group’s polygons created from the associated ethnicity and the provinces mentioned in the literature to be held by each group are remarkably similar.

EPRDF in Amhara

Left: Created on QGIS by Carly Gottorff. Right: Actual map of Ethiopia, source: http://reliefweb.int/sites/reliefweb.int/files/resources/3C4E623149B60FDCC12577580032E777-Map.pdf.

In the case above, the brown/orange polygon to the left are the ethnic areas associated with the EPRDF (Ethiopian People’s Revolutionary Democratic Front). On the right is an actual map of Ethiopia with the Amhara region highlighted, where the EPRDF was known to have the majority of its control. The EPRDF was located in central Ethiopia, corroborating the fight for independence as they attacked the capital. Specifically, the EPRDF was concentrated around the city Dessie in the Amhara region but they had wide control of the country, with a force estimated to be 70,000 by 1991 (de Waal, 1991, 308).

Oromo Liberation Front (OLF) in Oromia, Ethiopia

Left: Created in QGIS by Carly Gottorff. Right: Actual Map of Ethiopia, source:

The OLF (Oromo Liberation Front) was an ally to the EPRDF and helped them take the capital in 1991, maintaining control of the Oromia region of the country (de Waal, 1991). On the left is the map created in GIS, showing the light blue region as the ethnic areas associated with the OLF. On the right is an actual map of the Oromia region, pictured as green.

Eritrean People’s Liberation Front (EPLF) in the Northern Red Sea Region

Left: Created in QGIS by Carly Gottorff. Right: Actual Map of Eritrea, source: http://reliefweb.int/sites/reliefweb.int/files/resources/8171871E6E41D7B3852572210079083D-icrc_OPR_eri010706.pdf.

Lastly, the EPLF (Eritrean People’s Liberation Front) whose home base was located in the city of Nakfa in the Northern Red Sea Region (de Waal, 1991, 114). On the left the ethnic areas associated with the EPLF are shown in pink, on the right the map pictures the Northern Red Sea Region. In this last case the ethnic areas span a lot farther than the area emphasized in the literature, however because Eritrea succeeds this makes a bit more sense.

In all three Ethiopian cases, each ethnicity tied to a respective rebel group is mapped with relative accuracy to the territory bases accentuated in the literature. The other fifteen randomly selected ceasefires chosen to be examined were similar in their relative accuracy. It can therefore be concluded that ethnicity may be used as a relatively accurate proxy for rebel held territory.

Although studying civil conflict, and especially rebel groups can be extraordinarily difficult due to their nature, both can be analyzed through creative practices. Using ethnicity as a proxy for rebel held territory can be useful for further study of rebel groups outside of this research. The chaotic and often inhumane nature of civil conflict is emphasized in the destruction of Ben Tre, Vietnam. However, more complete data on the nature and location of rebel groups could help avoid analogous side effects of guerrilla warfare in future circumstances.


Carly Gottorff (gottorff@mit.edu) is a rising senior at SUNY Buffalo majoring in Political Science with minors in Economics and Statistics. She worked as a research intern in GOV/AB throughout the summer of 2016 through the MIT Summer Research Program (MSRP). MSRP is dedicated to helping lower income, minority and first generation students experience graduate school and receive valuable experience for this pursuit.

Works Cited

De Waal, A. (1991). Evil Days: 30 Years of War and Famine in Ethiopia. New York: Human Rights Watch

Narang, N. (2014). “Assisting Uncertainty: How Humanitarian Aid Can Inadvertently Prolong Civil War.” Int Stud Q International Studies Quarterly 59.1, 184-95.

Peace Research Institute Oslo. (2016). “Trends in Armed Conflict, 1946-2014.” Available from: http://file.prio.no/publication_files/prio/Gates,%20Nyg%C3%A5rd,%20Strand,%20Urdal%20-%20Trends%20in%20Armed%20Conflict,%20Conflict%20Trends%201-2016.pdf.

Taslitz, Andrew. (2011). “Destroying the Village to Save It: The Warfare Analogy (or Dis-analogy?) And the Moral Imperative to Address Collateral Consequences.” Howard Law Journal