Lethal Violence, Crime and State Formation in Cambodia
Australian and New Zealand Journal of Criminology Volume 35  2002:1-26
This paper estimates homicide rates and describes the nature of violence and crime for Cambodia. Limited data allows only a partial picture of the trends and nature of lethal violence. Post-war economic adversity combined with a weak state and underdeveloped “legal culture” contributed to an elevated rate of homicide. Frequent acts of murder-robbery, mayhem, political violence and banditry present a major threat to social and economic development. A murder incident rate of approximately 5.7 per 100,000 but a homicide rate of 9.3 per 100,000 was estimated for 1996, higher than most countries in the region except the Philippines. Political and economic adversity drove the homicide rate to 11.6 per 100,000 in 1998 similar to levels reached during 1993 the year of the first national elections. Usually homicides were between males and commonly arose from robbery, disputes and quarrels with most deaths resulting from gunfire. Extra-judicial death arising from police or “mob” actions accounted for high rates of suspect death and contributed significantly to the homicide rate. Rates of violent crime were higher in rural areas but Phnom Penh experienced high levels of property crime compared to the provinces. The homicide rate is compared with neighbouring countries and the role of modernisation, policing and crime discussed.
An apocryphal picture of the Cambodian countryside shows a signpost on a roadside bend outside of Phnom Penh which in Khmer states “Please do not dump body in this padi-field”. Cambodia’s reputation as one of the most lethal places in the world is perhaps justly earned. However, a decade after the 1991 Paris Accords, the departure of the United Nations Transitional Authority in Cambodia [UNTAC] and the establishment of the Royal Government of Cambodia [RGC] how violent is Cambodia? How does lethal violence in Cambodia compare with other Association of South East Asian Nations [ASEAN], developed and developing countries? Is Cambodia’s violent image justified?2
The combination of an armed society, the stress of post-war reconstruction and the impact of revolutionary genocide, poverty, fragmented institutions and a weak “rule of law” state produced more acts of mayhem, extra-judicial homicide and murder robbery than elsewhere. These forms of homicide exceeded homicide between intimates. The level of violence in Cambodia may not exceed that of other countries but it is certainly more lethal attaining apotheosis. The nature of lethal violence is a mixture of the ordinary and the bizarre. The following, Khmer newspaper accounts illustrate3:
April 23 Phnom Penh: Oul Phally, 33, was killed by robbers who stole her motorcycle on Street 318, Chamcarmon district. They shot her when she tried to speed up her motorcycle to get away. The offenders escaped with the victim's motorcycle
April 22 Kampong Cham Province: One person was killed and 32 wounded when a grenade exploded in the dancing place in Kok Sro Lauv village. So Van, believed to be the man who dropped the grenade, was killed in the blast. The source said that Van was very drunk and took the grenade from his waist and held it in his hand and certainly the grenade dropped and exploded
April 30: Two robbers were killed by a mob of people and the militia at Choum Chav precinct on the outskirts of Phnom Penh after they robbed a moto-taxi driver who was hired from Phnom Prosith. The source said the robbers hit the moto-taxi driver's neck and took the moto and sped away. But suddenly an angry crowd appeared because they heard the cries of the victim.
May 17: Hem Chandara, a robber was killed by military police after they brought him to show where his accomplice was, at Sangkat Toul Sang Ke, Khan Toul Kok. A source said he jumped from the car and tried to escape but military police shot to threaten him but unluckily the bullets hit the offender.
May 30 Phnom Penh: Ngweng Minh, 27, a Vietnamese man was killed after a group of moto-taxi drivers stopped him and beat him after he robbed one moto-taxi on Kampuchea Krom Blvd.
June 18 Kampong Cham: A commune policeman, Seng Kheavuth, was shot dead while he was swinging his son to sleep in his house in Tongrong commune, Prey Chhor district. Three unidentified men shot him with an AK-47 from under his wooden house and then escaped. A commune authority said that this was a revenge crime as there was nothing stolen.
July 4 Phnom Penh: Neang Sophea, 25, was chopped on his neck then shot to death by his girlfriend, Heng Sochea, in Kandal market. Sochea, 27, then shot herself to death with the same handgun. A letter found in her pocket by police said that she fell in love with Sophea on Nov 15, 1996, and later learned that he escaped from her to have another girl.
Aug 2: The body of an unidentified man was found in Koki Thom commune, Kien Svay district, Kandal. A newspaper reported that the victim had been killed and dumped at the site within the past few days because the body was very swollen. Investigations were continuing.
Sept 16: A body of an unidentified man was found in a pond in Phnom Penh Thmei precinct, Russei Keo district by a fisherman. The victim was shot and weighted with stone and dumped into the pond.
This paper estimates the prevalence and describes the nature of homicide in the Kingdom of Cambodia. An introductory attempt is made to theorise about the nature of violence in the context of the formation of crucial state capacities and institutions and the simultaneous modernisation of social and economic relations.
As a “re-organisation episode”, Cambodia offers an opportunity to observe the nature of crime following a state induced “collective disaster”, market reforms4 and the transformation of socioeconomic relations (Evans, Rueschemeyer & Skocpol 1985:364). The political settlement that led to the re-creation of a constitutional monarchy in 1993 also required the formation of a new state whose law and order institutions did not enjoy public support and legitimacy5. The absence of an established legal culture also restricted the ability of the new state to competently assert a consistent and credible legal threat to crime, illegality and disorder. Nevertheless, Cambodia is neither lawless nor are its policing institutions totally without any general deterrent effect. The Cambodian experience may also help to assess the general relevance of policing institutions in suppressing crime and producing order. Of theoretical interest is the nature of crime engendered by weak state institutions in the context of strong communal and non-state forces. Cambodia parenthetically reveals conditions of pre-modernity in the context of post-industrialisation that perhaps the advanced state may resemble stripped of the apparatus and symbolism of the "rule of law". When state regulation is weak, and private and customary regulation strong, what sorts of crime prosper?
Cambodia is still in the crucial stages of state formation following decades of war, genocide, insurrection and widespread civil dislocation and experiences crime and homicides different to that found in established states. Is crime, especially homicide, a sensitive barometer of social change and good governance? Does the level of violence reflect the stresses of a shattered economy and weakened human capital or reflect a culture of impunity based on an enduring “Realpolitik” of violence created by decades of conflict? Cambodia offers an opportunity to observe how in the process of state formation the establishment of a criminal justice system impacts on the nature and volume of crime. However, to address these questions we must first take into account the legacy of war and revolution.
Archer and Gartner (1976:960) in a seminal review of the effects of war on homicide rates demonstrated that substantial increases were observed [irrespective of the outcome of the conflict], and that high combat losses predicted higher levels of post-war homicide. They tested several explanatory models and concluded that a “Legitimation of Violence Model” that predicted an increase in homicide “…as a result of the pervasive war-time presence of officially sanctioned killing” was pervasive. Other explanations such as the “Social Solidarity” and “Cartharsis” or war fatigue models that predicted decreases or returns to normal levels of homicide were rejected. Alternative explanations for higher post-war homicide rates arising from distressed post-war economies, return of violent veterans or demographic changes also failed to account for the increases observed. Our ability to test these explanations for Cambodia is necessarily limited because of the destruction of almost all government records during the revolutionary period. Consequently, no reliable crime data exists pre-war [before 1970] and what is available is poorly defined and ambiguous. French colonial records are also incomplete but suggest that crime, especially homicide, was extremely rare during the protectorate period [1902-1954]. This appeared to remain so until the mid-1960s when civil disorder and mounting pressure from the Vietnam War engulfed ‘peaceful’ Cambodia 6.
Despite the scant picture of pre-war crime, Cambodia has been exposed to an extended ‘legitimation of violence’. We may argue that Archer and Gartner’s thesis, that war (and revolutionary) violence legitimated by the state carried into higher peacetime homicide. Indeed the very high rate of suspect/offender death may better reflect this process of violence legitimisation and, is in accord with many Khmer idioms on violent death (Chandler 1993; Keirnan 1996). Police and citizens alike commonly remark: “in Pol Pot time life was very cheap”; “criminals are merciless and must be killed”; “we must save ourselves because they [the state] are corrupt” or; “people are the law and must have justice”. Therefore, lawlessness and vigilantes are attributed by Khmer to the brutalisation experienced during the civil war [1970-1975], revolutionary period [1975-79] and occupation by Vietnam [1979-1989]. However, the processes of state formation and modernisation also play an important role in shaping the nature of crime and the form of homicide7.
Theories of Crime, Modernisation and Development
We have some theoretical guidance about the nature of crime as it might develop in a post-war and post-revolutionary developing nation such as Cambodia. Theories of modernisation, depending on the specific phase and locality, suggest that development will reduce violent crime but also increase property and other crimes (notably those against the state) as the rationalisation of modern governance is achieved. Modernisation requires a shift in economic modes of production from feudal/mercantilist to industrial forms or from "Asiatic" or command to market economies. This shift in productive forces produced in the classic form8 greater individualism, a significant middle class, weakened communal regulation, changed the nature of relationships from hierarchical to exchange and shifted social control from informal to formal modes. Also because of the civilising (or socialisation) effects of modernisation9 violence becomes more problematic and is subject to an intensified criminalisation process combined with an increasing reliance on bureaucratic surveillance and special policing institutions. Thus, in this functionalist version, modernisation in its early phases generates acquisitive crime by weakening social control and unleashing expectations10. In late modernisation, violence increases as conflict re-emerges due to rapidly changing modes of production and the fragmentation of post-modern identities.
Neopolitan (1997:360) combining Durkheim and Elias, argues violence declines because modernisation results in “…increased social equality and organic solidarity and a resultant civilizing of the human personality”. Modernisation or development is usually measured by economic development and, following the European experience, the degree of urbanisation or differences between urban and rural life. Cambodia at $US 270 GNP per capita is one of the least economically developed and urbanised countries in the region [see Table 5]. Cultural integration or homogeneity often associated with Durkheim’s idea of “organic solidarity” and measured by fidelity to religious, linguistic and customary beliefs has also been seen as influencing the amount of violent crime. Studies of the relationship between culture and homicide, using measures of religious and ethnic homogeneity, have not found evidence of a consistent pattern. Nor has a relationship between economic development and homicide been found (del Frate, 1998, Newman 1999, LaFree 1999)11.
Other explanatory theories of cross-national differences in violence include opportunity/stress theories that highlight changes in the pool of potential offenders and opportunities for violence that vary according to economic and social hardships. Thus the relative size of high risk groups (young unemployed males), population density, household size, income inequality, unemployment and infant mortality have all been employed with varying success as determinants of violence. Cambodia with its exceptionally youthful population, acute levels of income inequality and relatively high under-employment can also be vulnerable to the play of these hardships and their presumed deleterious effects on the risks of violence. Competition over scarce resources in the context of desperation can also produce greater recklessness that may be reflected in the high levels of robbery murder observed for Cambodia (Daly and Wilson 1999).
A concordant “civilising process” generated by modernisation reduces the incidence of violence as sensibilities about suffering increase and the demand for “blood” sacrifice decrease. Thus in modernising societies, violent crime will decrease and homicide rates will decline. The role of the ‘rational’ state through its security [monopolisation of violence], management and welfare functions is crucial in this civilising process. The state by improving health services, food/livelihood security and literacy, and suppressing crime legitimates its governance12. The RGC’s national development plan gives priority to the establishment of the institutional means to create a modern “rule of law” state by strengthening regulatory agencies, especially police and courts. This priority is supported by donor nations and the International Monetary Fund’s [IMF] rehabilitation program for Cambodia.
Theoretically, it may be useful to consider the extent to which the processes of modernisation influences the nature and extent of crime emerging from extreme intra-state conflicts. Can crime in Cambodia be characterised as pre-modern and does economic development suppress or exaggerate conflicts? Certainly higher levels of violence in the countryside than in the city but more property crime in the city than the country was observed. In accord with modernisation theory differences in urban and rural crime suggests a relationship between the strength of the state and opportunities for crime concomitant with development. Sporadic episodes of banditry, robbery and abduction are also reported more frequently in rural districts and these incidents reflect the relative weakness of the state and the limited radius of its policing institutions. This pre-modern picture is reinforced by the low enforcement capability and adherence to due process. The rehabilitation of legal processes after the almost complete destruction of the judicial corps has been in progress but is hindered by the absence of a tradition of rule by law and the low pay of state agents. As Chandler (1992) has shown these weaknesses were common in pre-revolutionary Cambodia and current political friction and abuses have precedent in the 1954 post-colonial kingdom.
Post-war economic stresses would also generate increases in both property and personal crime, especially in the city were wage labour is vulnerable to shifts in trade and investment. Although Cambodia does not have a critical food security problem the bulk of the population is rice dependent. Up to half of village, income was used for rice consumption that provided 80% of calorie intake. Rural poverty is a serious problem with as many as 40% or more living below the poverty line and there is endemic but moderate levels of malnutrition. Credit problems and indebtedness are widespread (Murshid 1998). Despite an abundance of land a land-less or land poor class has emerged and has become a source of accelerated migration to the city or towns. At the same time as the gap between rich and poor increases access to traditional sources of succour the rivers and forests (common lands) have become more restricted by large commercial interests. Thus, the revival of plantation economies and indentured work contributed to further pressure on the land-less. Cambodia is thus an economy dependent on subsistence rice agriculture, exploitation of forest and international aid but with the burden of massive post-war reconstruction. GDP growth was slashed to below 2% in 1997-1998 from 6.5% in 1996 due to the twin influences of the Asian financial crisis and the July 5-6 1997 coup which cut-off investment and all but essential aid. Already one of the poorest countries in the ASEAN group the impact of these economic difficulties adversely impacted on efforts to contain disorder and crime. Annual inflation reached 15% in the first half of 1998 placing considerable pressure on the poor and labouring classes as basic food and housing prices sharply increased while income declined. While GDP growth has recovered to 4.5-5.0% in 1999-2000 investment remains stalled and recovery further hampered by severe floods in 2000 (Cambodia Development Review 1998; Far Eastern Economic Review, December 14, 2000:87). Such crushing poverty contributed to intermittent banditry and placed already vulnerable groups under additional pressure. Crime victimisation is often catastrophic and may lead to impoverishment.
The establishment of rational, rule of law government is considered a basic pre-condition for the stable and predictable development of capital, markets and wealth (World Bank 1997). In developing economies, the transformation of the economy from feudal/subsistence forms to a post-revolutionary market economy requires a professional and predictable government. The reliability of law and policing institutions are seen as essential factors in the development of a market economy based on trade and commerce. Therefore, the effectiveness of legal institutions in creating stability and order is an essential test of the establishment of legitimate governance in the Kingdom13. Thus, the extent that police serve a general order and provide civil protection to the populace rather than serve a specific regime, the degree of lethal violence, may provide a measure of state strength. High levels of lethal violence, especially extra-judicial violence, are indicative of the states debility. “Many human rights critiques…fail to recognise police shortcomings as an expression of state weakness rather than of its strength” [Goldsmith 2001:18]. Given the violence threatened by powerful non-state actors, policing institutions are vulnerable to adopt partial and repressive measures that ultimately undermine their legitimacy14. In the context of conflict, the formation of autonomous state policing institutions is also uncertain because of underlying competition between various factions to provide security. Different policing agencies and factions within them thus compete to monopolise the “…universal sinews of state power” (Skocpol 1985:16), revenues and coercive force, and consequently reinforce political fragility.
Policing, Law and the Reformation of the Cambodian State
The fragmentation of the original civil society by war and genocide has required a conscious effort to re-invent and reproduce the social and moral order of an ideal pre-war Khmer nation. The 1991 Paris Accords reconstituted, under UNTAC supervision, the RGC as a pluralistic state made up of the former SOC, Khmer Rouge and Royalists under the rubric of “Nation, Religion, King”. A context in which the crucial struggle for the rule of law demands the new state’s nascent legal institutions monopolise violence and install nationwide means to resolve disputes and deal with criminals (Sok and Sarin 1998; Fernando 1998). However, an aid dependent RGC negotiates the means to effect state legitimisation while under the scrutiny of numerous non-government organisations [NGOs] notably the Cambodian Office of the United Nations High Commissioner for Human Rights [UNCOHCHR].
The rationalisation of governance is an ongoing process, structures are in place but variable means exist to assert the control of state institutions and its agents independently of regime politics and traditional patron/client relations. Until recently, the state has depended on military and quasi-military forms of governance albeit fractured along political/patronage and provincial lines. In the crucial areas of policing and dispute resolution, few resources are available and Cambodia remains vulnerable to organised crime activity15. Ample anecdotal information suggests that Cambodia has become a haven for eco-exploitation, drug and sex trafficking16, money laundering, small arms smuggling and other cross-border illegal activities. Cambodia has resembled the ideal de-regulated state and therefore an unfettered market for crime as well as “market solutions to insecurity” (Goldsmith 2001:12).
The transformation of the governance of state security from a military to civil form is an important goal of the national development plan along with improving food and personal security, human rights, health education and economic development. The shift required from a military command economy to a market economy is burdened by the 47% of the RGC revenue absorbed by the security forces [Ministries of Interior and Defence]. As little as 10% of meagre revenues went to education, 6% to health and 3% to agricultural development (Konrad Adenauer Foundation cited in Phnom Penh Post, Vol. 7, No. 17, 1998). Under the 1993 UNTAC de-mobilisation plan, large numbers of former soldiers were placed in policing roles on below subsistence wages. Consequently, the management and discipline of such a large body of armed and ill trained ‘police’ has been a major problem and source of impunity. Progressive reductions in under-employed untrained police and military planned by the year 2001 require numbers of police and military personnel to fall to 67,000 from the current 137,000. However, reductions in security forces and weapon availability have been hampered by factional differences in the key Ministry of the Interior [MOI] and the overall civilianising process will be lengthy and involved. Nevertheless, extensive training of police and military in constitutional responsibilities and the scrutiny of a large number of human rights NGO’s support a rule of law culture. These efforts contributed to the satisfactory performance of the security forces in the 1998 July election while under the command of the National Election Commission17.
The RGC operates in the context of a politically divided government of former antagonists - royalists, revolutionaries and socialists. Consequently state institutions are complex and open to factional conflicts and an effective law-making consensual process has not been achieved. Historically the Kingdom had a weak indigenous bureaucracy based on a French and Vietnamese colonial legacy in which recourse to state and legal institutions was grossly underdeveloped. Traditionally the Cambodian state was hierarchical with the revered King, the elite government officials18, the villager of Kompong and Preih [forest] and the sanga or Buddhist monks all in their place in the complex web of patronage and power. In crucial ways these traditional, often unmediated hierarchical relations found there expression in the utopian self-sufficiency program of the revolutionary Democratic Republic Kampuchea [1975-79] and continue to shape Cambodian personal and social relationships (Chandler 1992:53-54).
The ceaseless cycles of rice growing that define Khmer village life impose a customary order insulated even from the special, if relatively brief, traumas of the Khmer Rouge revolutionary utopia and the first Vietnamese invasion and occupation of modern times. In the context of the New Kingdom, these village continuities provide a natural social order replete with communal surveillance and control that reduce the need for intervention by the state in interpersonal disputes. Accordingly, the role of the state in creating order is limited to the extent that it may impose uniformity on these politically fragmented pre-existing social identities and relationships in rural Cambodia. A task perhaps never previously achieved by any Khmer state with the possible exception of the Angorian period.
The sanga, along with intellectuals, were the targets of special oppression by the Khmer Rouge because as carriers of traditional morality they were a source of resistance to the new order (Keirnan 1996). Buddhism was also denigrated as inimical to communist values during the Vietnamese occupation but, along with the King, it is promoted as a foundation of the new nation. The moral vacuum created by the dissolution of the post-colonial state and the failure of revolutionary idealism enabled the rapid return of this traditionally important institution. The sanga have revived their tenure and everywhere Pagodas have been re-built in this devout Buddhist society. The sanga and the kharmic laws have a protective, if poorly understood, role in shaping violence and crime. The re-establishment of Buddhism has been significant in the regeneration of indigenous moral order but its revival is challenged by modernity, materialism and new forms of crime.
The arch-criticism of Buddhism has been that it produced fatalism, conservatism and carelessness towards death (after all an opportunity for re-birth and enlightenment) as a by-product. These artefacts of Buddhism are often assumed to lead to a diminution of the sanctity of life in Buddhist societies in much the same way Catholicism is supposed to liberate the passions through forgiveness and the confessional. However, this neglects the emphasis on compassion and enlightenment in Buddhism that heightens the importance of the reciprocal nature of patron-client relationships so characteristic of Cambodian political and social life19. The Buddhist ethical system through stress on the four noble truths (which emphasise enlightenment, compassion and prohibit killing, stealing, lying and adultery), the middle way and the eight-fold path socialises a cooperative and docile human nature. The emphasis on avoidance of suffering and the accumulation of merit provides a potent traditional source of natural or internalised forms of social control against violence. Thus, a recognisable basis for the establishment of clearly defined laws is embedded into the culture. However, law making has been restrained by political instability and administrative fragmentation that reflects the continued importance of the Strok or district in national affairs. The ensuing hiatus allowed for ample regional and political differences in the creation and application of national law.
Although a new penal code is being drafted, current laws are based on the1992 UNTAC criminal code. This interim code failed to provide a comprehensive ethical system or secure due process consistent with Khmer values further eroding the legitimacy of law. UN sponsored efforts to rapidly develop a modern court and dispute settlement process included a “Judicial Mentor Program” that relied on guidance by overseas judges. This attempt to provide training for court officials [most former SOC judges] faltered because the poorly paid judges were prone to corruption or intimidation and access to courts prohibitively expensive to all but the elite. It is known that a substantial number of offences and disputes are resolved without the involvement of provincial or national courts. Therefore, some serious crime and much petty crime and disorder somewhat, as everywhere, will be under-reported20.
In the following, reported crime, especially lethal violence is described from available sources. The principal source of data is a report for the period 1992-1996 prepared by the Judicial Police Centre, MOI in July 1998 and subsequent summary annual reports. This source provides limited trend data covering the arrival of UNTAC and the establishment of the RGC to the end of 1999. In addition reports of NGO’s, Khmer newspapers, as well as interviews, case notes and field studies supplement official sources21. The focus shifts to the integrity of investigations into homicide and other serious crime and their relevance to the legitimisation process in the new state.
Law enforcement in Cambodia involves a number of different policing agencies including military police and it is unlikely crime records compiled by the Judicial Police Centre are complete. Because of uncertain counting and variable reporting behaviour, the most reliable approach to the interpretation of trends for most crimes in Table 2 is that they reflect the activities of police rather than “real” fluctuations in crime. Nevertheless for homicide we have several sources and may be more confident of the recording of these events.
Table 1: Crime Trends Recorded by Judicial Police 1992-1999
1992 1993 1994 1995 1996 1997 1998 1999
Murder 429 599 303 397 542 317 793 581
Armed Robbery 1414 1613 905 832 1345 887 1822 1396
Grenade 40 157 79 27 54 46 68 42
Rape 106 43 39 84 122 46 130 165
Kidnap 9 93 133 24 44 23 130 91
Poisons aa 14 47 42 23 24 21 0 20
Patrimonyb 8 6 20 12 6 n/a n/a n/a
Theft 1420 792 835 896 1471 868 1871 1789
Assault/Disputesc 515 267 353 423 1050 445 1114 1058
Fraud/pick-pocketd 23 117 163 214 233 97 244 248
Illegal Weapon 136 514 235 310 79 n/a n/a n/a
Other Offences n/a n/a 20 n/a 668 208 950 641
All Crime 4114 4248 3031 3260 5638e 2958 7122 6031
Rate 45.2 44.7 30.9 32.6 59.6 27.9 65.8 53.8
Sources 1992-1996 and 1997-1999 annual returns MOI Judicial Police Centre; data for 1997 incomplete. Notes: n/a = not available; (a) poisoning and kidnapping are presumed non-fatal; (b) comprises theft of cultural heritage; (c) records only injurious assaults; (d) offences combined in original source; and (e) total includes 201 injuries associated with grenade attacks.
Overall, the reported crime rate for all offences is very low in Cambodia with a mere 5638 offences recorded in 1996 or a rate of 59.6 per 100,000 population. This can be compared to the1253 offences per 100,000 recorded in Hong Kong, and the 110 per 100,000 recorded for the Philippines (see Table 6). The MOI report also attempts to take into account the effectiveness of policing by recording the number of offences “suppressed” [cleared or “solved”] and the number of offenders held in custody. Recent data is not available but for 1992-1996 between 28% - 36% of crimes reported by the MOI were suppressed. The number of new offenders imprisoned also fluctuated accordingly from 1045 and 2170. However, prison census data for 1995-1997 show the numbers held in prison increased from 2490 to 2909. These figures yield an estimated imprisonment rate of 26-29 per 100,000 and are comparable to rates reported for the Philippines but considerably lower than the 200 prisoners per 100,000 in Hong Kong22.
Although the conventional definition of homicide as murder and non-negligent manslaughter is adopted the sources necessitate ambiguity in defining homicide because of the unclear status of justifiable homicides and other forms of non-negligent manslaughter recorded by National Judicial Police. The actual number of murder victims was not reported and must be estimated from previous years, deaths recorded under "losses" by the MOI, field studies and newspaper reports. A detailed estimate of the homicide rate is provided below and only a rate for murder events can be utilised to compare risks of lethal violence between the city and countryside. During 1996, 542 murder events were recorded at a rate of 5.7 per 100,00023 but the rate varied substantially by province. For the city of Phnom Penh it was 4.7 events per 100,000, 6.4 for the densely populated central province of Kompong Cham and in Seim Reap province, site of the famous Angor temples it was13.8 per 100,00024.