Khmer Rouge Victim Numbers, Estimating

There are at least four possible approaches to determining the number of people killed in a given instance of genocide or other crimes against humanity. The best estimate of such fatalities is possible if the perpetrators have kept accurate records, but this seems rare. A second approach requires the investigation of mass graves, either by taking an actual count of exhumed bodies, or making an estimate based on the number and the sizes of the graves. A third approach is the demographic analysis of census data or other population data. A fourth approach involves interviewing survivors of the violence about fatalities in their families, followed, in most cases, by a statistical extrapolation from the results of the interviews. The advantages and disadvantages of each approach can be understood by reference to the case of Cambodia, where all of these methods have been applied.

One of the first attempts to gauge the magnitude of the Cambodian genocide was a project undertaken by the People's Republic of Kampuchea, which took control of the country when the Khmer Rouge fell from power. In a four-year project, the "Research Committee on Pol Pot's Genocidal Regime" conducted survivor interviews, along with mass grave exhumation and analysis, to come up with the figure of 3.316 million dead during the Khmer Rouge period (1975–1979). Later analysts have questioned the methodology used in making this estimate, arguing that its results were likely inflated by the double counting of some victims—for example, a victim reported as killed by a family member may have been counted again when the body was exhumed from a mass grave—along with an underestimate of net migration.

Another early effort at estimating the magnitude of the Khmer Rouge genocide was published in 1980 by the U.S. Central Intelligence Agency (CIA), which employed demographic analysis. The CIA began with Cambodia's 1962 census, then used subjective reports to estimate trends in fertility, mortality, and migration through 1979. The agency calculated that Cambodia's population had declined by somewhere between 1.2 and 1.8 million during the Khmer Rouge regime. Numerous assumptions underlying this analysis have been criticized, particularly its conclusion that the number of rural dwellers increased marginally between 1975 and 1979. Later analysts determined that there was significant excess mortality among the peasant population, and by implication, a relatively higher overall death toll.

More recent demographic analyses have taken advantage of post-genocide population data to refine the CIA's estimate. Based on data collected through a Cambodian administrative census conducted in 1980, Judith Bannister and Paige Johnson calculated a population loss between 1975 and 1979 of 1.8 million. In their 1993 report, they concluded that 1.05 million of these deaths were excess mortality. Patrick Heuveline employed birth cohort data derived from the 1993 electoral register to determine that the most likely figure for excess mortality during the Khmer Rouge regime was 2.2 million, also concluding that about half of these deaths, or 1.1 million, were from violent causes, primarily execution. All of the well-understood weaknesses of census and other population data are imported into such analyses, particularly with methodologically unsound censuses such as the 1980 count. This inherent propensity for error is further magnified by the assumptions made to compensate for missing data, such as fertility rates.

Interview and survey data have also been used to construct estimates of the death toll during the Khmer Rouge genocide. Ben Kiernan launched one of the first such efforts, interviewing some 500 subjects in 1979 and 1980, and extrapolating his findings to the national population for an estimate of 1.5 million deaths. He later refined his estimate to 1.671 million. Similarly, Steve Heder surveyed more than 1,000 Cambodian subjects, concluding that there were approximately 1.7 million deaths under the Khmer Rouge, with a death rate of 33 percent among urban Cambodians, 25 percent among rural Cambodians, and 50 percent among Sino-Khmer. A more systematic interview project was conducted by Marek Sliwinski between 1989 and 1991, with some 1,300 respondents. His data yielded an estimate of 1.84 to 1.87 million excess deaths during the Khmer Rouge regime. It is notable that these three interview approaches yielded very similar results, ranging from 1.5 to about 1.9 million. Nonetheless, this method entails numerous potential sources of error. It is difficult to construct a representative and random sample of subjects. Moreover, this method also depends on estimates of pre- and post-genocide populations, which are typically unreliable. This method does, however, have the advantage that it can be carried out by a single investigator, relatively soon after the genocide has been halted.

A hard count of victim remains is yet another potential approach. Such a project has been underway at the Documentation Center of Cambodia since 1995. The effort involves mapping mass graves and estimating of the number of victims contained therein. As of May 2003, the Documentation Center had identified 19,471 mass graves, which were believed to contain the remains of an estimated 1.1 million victims of Khmer Rouge execution. Interestingly, this matches Heuveline's estimate of the number of deaths from violent causes, even though Heuveline reached his figure by a very different method. An advantage of the hard count method is that it is primarily empirical, and does not rely on overall population estimates. Nonetheless, error can be introduced from several sources, such as the method used to estimate the contents of graves. The possibility of faulty witness testimony regarding the origin of mass graves is also a problem. Finally, the hard count method cannot necessarily distinguish between excess mortality due to execution and deaths due to other causes, such as starvation, disease, and exhaustion.

The use of perpetrator records to determine the magnitude of a genocide has rarely, if ever, been implemented, because of the problem of gaps in recordkeeping. For example, at most of the 167 Khmer Rouge extermination sites identified in Cambodia, no contemporaneous records appear to have survived, if indeed they were maintained in the first place. Even at the most meticulously documented Khmer Rouge extermination center, Tuol Sleng Prison, gaps in the records have resulted in death toll estimates ranging from a low of 15,000 to a high of more than 20,000, quite a high degree of uncertainty. Obvious questions about the integrity of data produced by perpetrators also increases doubts about the reliability of this method.

The challenges apparent in all these varying approaches to estimating the magnitude of genocide or crimes against humanity suggest that analysts should approach this task with a certain degree of humility. Public records such as birth and death registers are typically among the first casualties during instances of extreme socio-political upheaval. This problem is often compounded by unreliable population data prior to and in the immediate aftermath of the crisis. Humans are notoriously unreliable as witnesses, and plumbing the depths of mass graves is a labor intensive, uncertain undertaking. The optimal approach may be to pursue all these methods—hard count, demographic analysis, and interview data—and, mindful of the pitfalls of each, triangulate the results into a range of estimates. In the Cambodian case, this range is from 1.7 million to 2.2 million, with the more recent and methodologically sophisticated efforts tending to produce results in the upper end of that range.

SEE ALSO Cambodia; Khmer Rouge; Khmer Rouge Prisons and Mass Graves; Pol Pot; Statistical Analysis

BIBLIOGRAPHY

Banister, Judith, and Paige Johnson (1993). "After the Nightmare: The Population of Cambodia." In Genocide and Democracy in Cambodia: The Khmer Rouge, the United Nations and the International Community, ed. Ben Kiernan. New Haven, Conn.: Yale University Southeast Asia Studies.

Central Intelligence Agency (1980). Kampuchea: A Demographic Catastrophe. Washington, D.C.: Author.

Documentation Center of Cambodia (May 2003). Master Genocide Site Data. Phnom Penh: Author.

Etcheson, Craig (2000). 'The Number': Quantifying Crimes Against Humanity in Cambodia. Phnom Penh: Documentation Center of Cambodia.

Heuveline, Patrick (1998). "'Between One and Three Million': Towards the Demographic Reconstruction of a Decade of Cambodian History (1970–1979)." Population Studies 52(1):49–65.

Heuveline, Patrick (2001). "The Demographic Analysis of Mortality in Cambodia." In Forced Migration and Mortality, eds. Holly E. Reed and Charles B. Keely. Washington, D.C.: National Academy Press.

Kiernan, Ben (1996). The Pol Pot Regime: Race, Power, and Genocide in Cambodia under the Khmer Rouge, 1975–1979. New Haven, Conn.: Yale University Press.

People's Republic of Kampuchea, Research Committee on Pol Pot's Genocidal Regime (1983). Report of the Research Committee on Pol Pot's Genocidal Regime. Phnom Penh: People's Republic of Kampuchea.

Sliwinski, Marek (1995). Le Génocide Khmer Rouge: un analyse démographique. Paris: Editions L'Harmattan.

Craig Etcheson