75057 Research and Dialogue Series: Paper No. 1 The Socio-economic Costs of Crime and Violence in Papua New Guinea Trends in Crime and Violence in Papua New Guinea Sadaf Lakhani and Alys M. Willman May 2014 Social Cohesion and Violence Prevention Team Social Development Department © 2014 The World Bank Group 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org/socialdevelopment E-mail: asksocial@worldbank.org All rights reserved. This paper has been prepared as part of a broader study to understand the socioeconomic costs of crime and violence to businesses, government agencies, Civil Society Organizations (CSOs) and households in PNG. This work was requested by the Prime Minister and was undertaken with extensive input from international partners and local stakeholders. The papers in this Research and Dialogue series are informal publications of the World Bank. They are circulated to encourage thought and discussion. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. Contents Acknowledgements ..................................................................................................................... v I. Introduction ............................................................................................................................... 1 II. Homicides appear to be stabilizing, with broad variation across provinces............... 2 III. Overall rates of crime and violence also appear to have stabilized, but again there is important variation between provinces and towns ..................................... 5 IV. Crime and violence are concentrated in particular ‘hot spots’ ..................................... 7 V. The types of crime and violence experienced vary significantly by region and by gender ............................................................................................................................... 9 VI. Men and women experience crime and violence differently, with gender- based violence the most common form of crime experienced by women ...................... 10 VII. A large proportion of criminal activity involves violence, and this appears to be increasing in certain areas .............................................................................................. 11 VIII. The use of firearms in violent crime is increasing..................................................... 14 IX. Little is known about perpetrators of crime and violence, other than that they tend to be male, and known to their victims ............................................................... 16 X. Arrest rates have remained consistent and low, even in years when reported crimes have decreased ............................................................................................................... 17 XI. Conclusion ............................................................................................................................ 19 Summary of findings .............................................................................................................................19 Gaps in data and knowledge ................................................................................................................20 Works cited.................................................................................................................................. 21 Annex 1: Data on crime and violence ..................................................................................... 24 1. Data sources and gaps .......................................................................................................................24 2. Challenges in classification of crime ................................................................................................26 Graphs and charts Graph 1: Murder rate PNG, years 2000–2010 .......................................................................................... 2 Graph 2: Comparison of murder rates from different data sets ........................................................... 3 Graph 3: Murder rates, by province, 2007 and 2010 ............................................................................... 4 Graph 4: Murder rates regional and worldwide 2010*.......................................................................... 4 Graph 5: Comparison of official crime statistics and victimization survey estimates, Lae 2005 ................................................................................................................................................................ 7 iii iv ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Pie charts 1-6: Violent crime is growing as a proportion of total crimes reported (2007 and 2010*) ............................................................................................................................................................ 12 Pie chart 7: Western Highlands, reported incidences of violent crimes, 2010 ................................... 13 Pie chart 8: NCD, reported incidences of violent crimes, 2010 ............................................................ 13 Graph 6: Number of firearms offenses reported nation-wide is increasing since 2008 .................... 15 Graph 7: Comparison of crime reported and arrests made (numbers), 2000-2010............................ 18 Graph 8: Reported crimes and arrests compared................................................................................... 19 Acknowledgements This publication has been peer reviewed by Keith Krausse (Small Arms Survey), Eric Kwa (Papua New Guinea Constitutional and Law Reform Commission), Alexandre Marc (World Bank, Centre for Conflict Security and Development), and Matthew Stephens (World Bank, Social Development). Special thanks go to Laura Bailey-World Bank, Country Manager, Papua New Guinea, for her guidance, and the team at the PNG country office for their logistical support. The team would also like to thank Almah Tararia (Consultant), Debbie Isser (World Bank, Justice for the Poor), Nick Menzies (World Bank, Justice for the Poor), Sakuntala Akmeemana (World Bank, Justice for the Poor,) Gary Milante (World Bank, OPCFC), and the team at the Law and Justice Sector Secretariat, Government of Papua New Guinea, for their feedback during research and preparation. v Trends in Crime and Violence in Papua New Guinea I. Introduction Crime and violence are widely viewed as posing a considerable challenge to development in Papua New Guinea (PNG). The full scale of crime and violence in PNG is difficult to assess, given the scarcity of national-level studies and a distinct urban bias in the available studies. Yet various commentators and surveys estimate that violence victimization rates in PNG are among the highest in the world (UNDSS 2005, Small Arms Survey 2006; Haley and Muggah 2006, Haley and May 2007, Goldman 2007, Barker n/d, Goldman 2007). This paper presents some preliminary findings regarding the prevalence of crime and violence in PNG. It was prepared as part of a broader study to understand the socioeconomic costs of crime and violence to businesses, government agencies, Civil Society Organizations (CSOs) and households in PNG. This work was requested by the Prime Minister and is being undertaken with extensive input from international partners and local stakeholders. This paper based on a review of literature and available data from official sources (primarily the Royal Papua New Guinea Constabulary (RPNGC) and from crime victimization surveys conducted in several urban areas of PNG by the Law and Justice Sector Secretariat (LJSS) 1, as well as on interviews collected during a World Bank Scoping Mission in December 2011. The paper was discussed with several governmental and non-governmental stakeholders in follow-up missions conducted in October 2012 and April 2013 and has been refined on the basis of those discussions. The different data sources reviewed and the most significant challenges with the data available are noted in Annex 1. The challenges in partial data and questions concerning the methodology used for collecting and collating some of the data sets and data integrity call for some caution in interpreting the findings, in particular making generalizations about the wide diversity of provincial experiences on the basis of geographically limited data sets. 1 The PNG Law and Justice Sector Secretariat (LJSS) was established by the National Coordinating Mechanism (NCM) to assist the formal agencies working in this sector. Among some of its responsibilities include guiding the sector and relevant agencies, as well as monitoring performance in the sector. 1 2 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea II. Homicides appear to be stabilizing, with broad variation across provinces Data on homicide—widely considered the most accurate indicator of overall levels of violence in a country—are notoriously problematic in PNG. National police (Royal Papua New Guinea Constabulary—RPNGC) data suggest a homicide rate of 10.4 per 100,000 people in 2010. The data shows extensive fluctuation from year to year, but with an overall increase in homicides in PNG since 2006 but no increase since 2000.(see Graph 1). Graph 1: Murder rate PNG, years 2000–2010 12 Homicie rate per 100,000 10 8 6 rate per 4 100,000 persons 2 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year Source: RPNGC crime summary statistics. There is a very large discrepancy between homicide estimates using official (RPNGC) data and those based on data from victimization surveys (see Graph 2: comparison of homicide rates from different data sets). While victimization surveys (which ask a sample of people about their experience with crime) generally tend to capture higher rates of crime than official data (which rely on crimes that have been reported) in most countries 2, the differences seen in the PNG data sets are noticeably large 3 2 See UNODC and UNECE (2010) for a full explanation of the difference between police recorded data-sets and data captured by crime victimization studies. 3 Failure to report homicides to police is an important factor in the discrepancy; however, based on in-country consultation the World Bank team found that it is likely that methodological difficulties of implementing crime victimization surveys in PNG may have contributed to the large difference. A revision of the crime-victimization survey instrument is has been planned by the LJSS, as such future victimization studies will be able to reveal to the extent of the difference attributable to methodological challenges. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 3 Other data sources, although often from a far smaller sample sizes, also point to large discrepancies between the estimated total homicides committed and those reported officially. Lewis (2007) and Haley and Muggah (2006) quote the Regional Administrator for western Southern Highlands Province reporting that there had been over 200 conflict-related deaths in Tari District alone in 2003–04, while the RPNGC data recorded only 40 homicides for the entire province during the same period. Corroborating data from hospital facilities and morgues country-wide has not been systematically collected, but data from a study of hospital records from Mendi (Winnington 2008) also indicates significant under-reporting of homicide. Consultations with the RPNGC and other stakeholders on the murder rates confirmed that there is likely to be significant under- reporting of murder to the RPNGC. The reasons for this are explored in Paper no.2. This significant under-reporting means that official data on murder rates is less reliable an indicator of overall crime rates than it is in other countries. Graph 2: Comparison of murder rates from different data sets 800 750 700 Victimization surveys (2008) Homicie rate per 100,000 600 RPNGC (2008) 500 400 323 300 200 100 66 33 0 Lae NCD Districts Source: RPNGC crime summary statistics and Victimization Studies for Lae and NCD (2008). There appears to be important variation in homicide rates across provinces and towns that is hidden by national-level averages (see graph 3). .According to RPNGC data, low homicide rates are found outside of the Highlands and the major urban areas. Lae and National Capital District 4 reported the highest homicide rates in the country (66 and 33 per 100,000 persons respectively in 2010). These are nine and three times the global average homicide rate (7 per 100,000) and classified as ‘high’ according to UNODC benchmarks 5. The homicide rates for Lae and NCD are the highest found in the East- Asia Pacific region, and are amongst the highest found world-wide 6. Graph 4 compares 4 National Capital District (NCD) comprises Port Moresby and its immediate surrounding areas. 5 UNODC (2011:20). Rates below 10 are considered ‘low’, between 10 and 20 are ‘Medium’ and over 20 are ‘High’. 6 Ibid. 4 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea the rates of Lae and NCD with selected cities in East Asia and elsewhere to put these numbers in international perspective. It shows that homicide rates in PNG are high compared with both regional and global comparators. Graph 3: Murder rates, by province, 2007 and 2010 90 Homicie rate per 100,000 80 70 2007 60 2010 50 40 30 20 10 0 Province Source: RPNGC crime summary statistics. * Referred to in the RPNGC crime data as ‘North Solomons’ or ‘Solomons’. Graph 4: Murder rates regional and worldwide 2010* 140 116.6 Homicie rate per 100,000 120 100 80 66 59.9 60 40.1 40 33 19.1 20 11.3 5.6 4 5.3 4 1.1 0.5 0 Region Sources: Data from UNODC, 2007¹ 2008² 2009 ³data, Lae and NCD rates are from RPNGC data. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 5 The RPNGC data also show large variations over time. For example, the homicide rate for Western Highlands in 2007 was 78 per 100,000 persons and 9 homicides per 100,000 persons in 2010. Large fluctuations between data for years 2007 and 2010 were also seen in Lae, the Southern Highlands and Enga, and to a lesser extent, in New Ireland. Again, these numbers should be taken with caution, as they could be a result of inconsistencies in data collection. III. Overall rates of crime and violence also appear to have stabilized, but again there is important variation between provinces and towns At the national level, crime rates appear to have stabilized since year 2000. Victimization surveys and qualitative work conducted in the mid-2000s suggested a high prevalence of crime and violence in PNG (c.f Haley and Muggah 2006, Haley and May 2007, Goldman 2007, Barker no date). More recent data from both the RPGNC and from victimization surveys suggest that crime has stabilized in recent years, and that in 2010 there was an overall reduction in total crime levels compared with those in 2000 (see table 1). Table 1: Total crimes reported, years 2000–2010 Year Total no. of crimes reported % change from previous year 2000 13,292 2001 12,043 -9.4% 2002 11,285 -6.3% 2003 11,813 4.7% 2004 5,221 -55.8% 2005 10,459 +100.3% 2006 6,540 -37.5% 2007 7,448 13.9% 2008 8,511 14.3% 2009 9,550 12.2% 2010 9,785 2.5% Source: Data from UNODC, 2007¹ 2008² 2009 ³data, Lae and NCD rates are from RPNGC data. 6 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Victimization surveys reveal different patterns in reported crime for different geographical areas. In the National Capital District (NCD), the percentage of households reporting victimization of at least one crime decreased from 67% in 2004 to 38% in 2009. Similar decreases were seen for households that were the victims of more than one crime, of multiple crimes, and of five or more repeat crimes (LJSS, NCD, 2009). Kopoko also saw a decrease in reported crime from 2007 to 2010, with the exception of assault (LJSS Kopoko 2010). In Lae, crime appears to have increased for 11 of the 12 categories of crime reported in 2010 (compared with 2005 and 2008 data), and the number of households that reported being victims of at least one crime stayed steady at 78% (LJSS Lae 2010). Data on overall crime rates suffer from the same inconsistencies across sources as those noted for homicide data. First and foremost, the two largest data sets on crime and violence-from the RPNGC and urban crime victimization surveys- are both likely to significantly under-report crime and violence overall, and there is important regional variation in this. In the victimization surveys citizens respond that they report less than half of all crimes to police, ranging, in 2005 from 33% in NCD to 15% in Arawa. The reasons for this range from poor access/ long distances to police stations to greater cultural relevancy of traditional dispute resolution mechanisms (see brief 2). The problem of underreporting in the data is larger for some crimes than others. In the victimization surveys, participants said they were more likely to report the “most troubling crimes,” 7 which were generally understood to be those that involved the highest economic costs (for example property theft, breaking and entering) (JAG 2008). Reporting rates for crimes other than those ‘most troubling’ respondents are likely then to be even lower. The literature also notes that gender-based violence such as rape, sexual assault and forms of domestic violence, is likely to be under-reported in both the official statistics and in the urban crime victimization surveys for PNG (Lewis et al 2007, Eves 2006, JAG 2008, LJSS Lae 2010, LJSS NCD 2010, Kopoko 2010) 8. In cases of sexual and domestic violence, fear, shame and the idea that domestic violence is both a family matter and tolerated prevent reporting. Other reasons for non-reporting of some crimes include the desire for seeking ‘compensation’ payments (outside of the formal justice system) to local conceptualization of what constitutes a crime. Petty theft , for example may often be seen as a nuisance rather than a crime (these are explored in more detail in 7 ‘Most troubling’ or concerning crimes, as identified by respondents of the survey, varied between urban area but included the following: stealing of property, use of firearms, breaking and entering into domestic property, damage to property and unprovoked violence. The victimization studies found that financial impact is what makes these crimes the ‘most troubling’ (JAG 2008). 8 Each of the urban crime victimization surveys notes that they “can only give indirect evidence at best of minimum levels of domestic violence and sexual assault cases” (see for example LJSS Lae 2010, p8). Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 7 Paper No. 2 on Drivers of Crime and Violence). Comparison of the data and our consultations with RPNGC revealed that overall, reported crimes may constitute only a very small percentage of all crime. For example the 2005 victimization survey conducted in Lae estimated that only 2.3% of all crime victimization documented by respondents was reported to the police (see graph 5, comparing official crime statistics with estimates from victimization surveys). Graph 5: Comparison of official crime statistics and victimization survey estimates, Lae 2005 80,000 Victimization surveys 68,849 70,000 Police crime reports 60,000 Number of crimes 50,000 40,119 40,000 28,730 30,000 20,000 10,000 1,597 746 851 0 property crime violent crime total Type of crime Source: JAG 2008. IV. Crime and violence are concentrated in particular ‘hot spots’ Crime and violence is not experienced equally across PNG, but is overwhelmingly concentrated in ‘hot-spots’. There is strong evidence that much crime is concentrated in certain geographical areas in PNG (C.f Eves 2006, Haley and Muggah 2006). Even within urban areas, many of which, the available data suggests, experience higher crime rates than rural areas, crime is disproportionately concentrated within certain quarters in the towns or cities 9 . ‘Hot spots’ for crime (i.e. those experiencing crime incidence consistently above the national average) identified through analysis of the RPNGC data in 2010 were Lae, NCD, Madang, East New Britain, West New Britain, the Western 9This concentration of crime in certain parts of urban areas is noted in the LJSS NCD (2009) and JAG (2008) 8 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Highlands, and Enga. Large gaps in the RPNGC data for several provinces means that it is difficult to determine where else there may be crime hot-spots. 10 In the ‘hot spots’, crime appears to be increasing, especially certain crimes. This runs contrary to national trends, which suggest a stabilization in crime rates overall. In Lae, for example, while overall crime decreased between 2005 and 2008, an overall increase in crime has been seen between 2008 and 2010. The 2010 victimization data indicates that aggregated property crimes in Lae have actually doubled since 2008, and aggregated violent crimes more than doubled in 2010 compared with 2008 (LJSS Lae, 2010). The 2010 Lae study notes that there have been substantial increases in eleven of the twelve crimes measured by the survey. It is important to note that only 23% of households in Lae had not experienced any crime in 2010. This figure has remained at a similar level for the three surveys conducted so far in Lae, but the number of households reporting multiple crime victimization 11 has grown. NCD in particular seems to exhibit this worrying trend, of the greater vulnerability of certain sections of the population to crime victimization. Haley and Muggah (2006) for example, found that crime is concentrated in the settlements in the capital. More recent victimization data for NCD (for 2009) also identified areas within NCD with concentrations of crime-the settlements however, were not amongst those areas identified. Lack of predictability is shown in year on year variations. The victimization study for NCD for 2009 notes for example that, compared with previous years’ findings, there was a lack of consistency in which areas experience the highest levels of crime victimization within NCD (LJSS NCD 2009). The NCD 2009 study also finds that crime was high even in more affluent areas, where more sophisticated security infrastructure and private security are in place. Renbo, an affluent suburb of NCD, with high security had the highest figures for car theft and firearm use of all of the survey sites within NCD (LJSS NCD 2009). Crime rates in some settlements are overlaid a growing trend of ‘gentrification’—where more wealthier young families and middle-class government workers are moving to settlements due to rising housing prices in Port Moresby. 10 The RPNGC crime summary statistics have large gaps in the data for many provinces and towns. For some—such as for the Southern Highlands Province, Enga and Morobe—data is absent for a limited number of crime categories. For other provinces—such as North Solomons and Gulf, and Manus (for year 2009)—data is missing for most of the crime categories. For each of the provinces and towns experiencing gaps in data, the categories of crime where the gap exist show some variation from year to year. 11 Multiple crime victimization is where the individual or household experiences more than one incidence of crime-be it the same crime or a different crime-during a given year. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 9 V. The types of crime and violence experienced vary significantly by region and by gender Robbery and assault are the most frequently reported crimes in PNG. According to both the official RPNGC data and the victimization surveys conducted in the three main urban areas in PNG, the most commonly reported crimes included robbery (the most commonly reported crime in all the ten years 2000-2010 according to the RPNGC data) and Grievous Bodily Harm. ‘Stealing’ was amongst the top three crimes in the official data, and ‘Stealing more than K1000’ was one of the most frequent crimes reported in the urban crime-victimization surveys. Other crimes with high incidences in the data sets include the use of firearms, vandalism and damage to property, sexual offenses including rape, murder, drug offenses and arson 12 (see Table 1). Table 1: Most frequently occurring crimes a RPNGC reporting (numbers reported 2010) Urban crime victimization studies (2009 and 2010) Robbery (1,321) Property stolen (more than K1000) GBH (1,291) Assault Stealing (1,083) Breaking and stealing Rape (1,023) Vehicle theft Murder (729) Unprovoked violence Sources: a. Data from victimization studies for Lae 2010 (LJSS Lae, 2010), Kopoko 2010 (LJSS Kopoko 2010), and NCD 2010 (LJSS, NCD 2010). Domestic break-ins are now more commonly reported than break-ins into commercial property. Property crime is identified by both the RPNGC data and the victimization surveys as constituting a large proportion of total crimes committed. Rates of break-ins to commercial properties in 2004 and 2005 and 2006 (disaggregated data prior to 2004 was not available) were roughly equal to incidences of breaking and entering into private property throughout the country. By 2010, and looking at the preliminary data for 2011, while overall rates of each were lower than in previous years, the trend has reversed, with domestic break-ins higher than those experienced by commercial 12Most frequently reported crimes in RPNGC for years 2000-2010, and most frequently reported in the urban crime victimization surveys Kopoko 2010 (LJSS Kopopko 2010) Lae 2010 (LJSS Lae 2010), NCD 2008 (LJSS NCD 2010), NCD 2009 (LJSS NCD 2009) and JAG (2008). 10 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea properties 13. The increasing use of private security firms by businesses may be a factor in this trend. Further investigation however is required to determine whether it is simply reporting trends that have changed, or whether actual types of crimes committed have changed, and if so, why. VI. Men and women experience crime and violence differently, with gender-based violence the most common form of crime experienced by women Men and women experience crime and violence differently, with gender-based violence likely to be the most common form experienced by women nationally. The urban crime victimization studies provide useful data on gendered experiences of crime, both as victims, and as perpetrators. The data largely points to very different forms of victimization for men and women. While stealing, the most common crime, was experienced almost equally by both sexes, other crimes had strong gender associations. For example the NCD survey for 2009 noted that women were five times more likely to be victimized at home than on the street. In nearly half of the most serious crimes, the victim knew the perpetrator. Women tended to be victims of men known to them: a relative, or neighbor. Men, on the other hand, were more likely to come into conflict with strangers outside the home. Alcohol was cited often by respondents as a factor in violent crimes experienced by men, as well as in those committed by men. Recent, qualitative analysis of gender-based violence suggests that victimization is not significantly different across women of different socio-economic status (Lewis et al 2007). More highly educated women, and those from higher-income households experience intimate partner violence (IPV) roughly as frequently as those with lower- education and from lower-income households. The age of the IPV victim and perpetrator were the only factors that did show variation in perpetration of IPV, with older women slightly more likely to experience this form of violence than younger women 14. A significant percentage of gender-based violence also affects young males. Whilst gender-based violence is often conflated with crimes committed against women, in fact, an important social group subjected to violence in PNG is young men under the age of 15 years (Eves 2006, UNHABITAT 2004) Sexual crimes, as a subset of gender-based violence- are committed against children of both sexes, and also young adults, including 13Data from RPNGC 14The assumption made here is that age gaps between partners are comparable for young and older women. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 11 males; the UNHABITAT survey of young people (UNHABITAT 2004) found that 44% of young sexual abuse victims are male. Gender-based violence is highly prevalent and total crime victimization for women is likely far higher than is reported or recorded in either the official data or by the victimization studies. While widely recognized in the literature as a serious and widespread problem (C.f. NSRRT & Jenkins 1994; UN HABITAT2004; Bradley 2001, MSF, No date, Eves 2010, Lewis et al 2007), up -to-date figures on the extent of domestic violence and other forms of gender-based violence in PNG, such as sexual violence, are difficult to identify. The two largest data sets available—those from the RPNGC and from the urban crime victimization surveys—are both likely to underestimate the extent of crime victimization experienced by women, as violence within the household is likely to be under-reported to both the police and in the victimization surveys. Our consultations found that considerable variation in prevalence is likely to exist between provinces, but that historical trends, linked to cultural traditions were undergoing social changes. VII. A large proportion of criminal activity involves violence, and this appears to be increasing in certain areas In general, crime in PNG is characterized by high levels of violence. A survey on young people and crime indicates that crimes in Port Moresby involved a high level of violence (38%) compared to other cities around the world, where a maximum 25-30% of crimes involve violence (UNHABITAT 2004). The same study found that young perpetrators of crime often reported the use of violence in conducting property crimes. The urban victimization studies also illustrate that violence was commonly used in property crimes 15. Analysis of official data conducted for this report suggests that in some parts of the country, the majority of crimes reported to the police involve the use of violence 16. National averages of the official data reveal a small increase in the proportion of crimes that are violent; from 47.2% in 2007 to 49.7% in 2010. However, national-level averages hide significant variation between provinces. World Bank analysis of the RPNGC data suggests that violent crime, as a proportion of all crime reported, has been growing even more quickly in certain parts of the country. Further analysis is necessary but the pattern correlates with the crime ‘hotspots’, i.e. provinces with high overall crime rates; including Lae, NDC, the Southern Highlands and Western Highlands, Enga, Madang, West New Britain and East New Britain (see pie charts 1-6). 15 See JAG 2008. 16 RPNGC data. 12 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Pie charts 1-6: Violent crime is growing as a proportion of total crimes reported (2007 and 2010*) NCD 2007 Western highlands 2007 Lae 2007 44.6 47.6 42.7 55.4 52.4 57.3 NCD 2010 Western highlands 2010 Lae 2010 Violent crimes 41.1 46.8 49.7 50.3 53.2 58.9 Other crimes Source: RPNGC data for 2007 and 2010. Note: *Violent crimes: Murder, manslaughter, rape, other sexual offenses, robbery, robbery MV, firearms, abduction. Other crimes: all other crimes reported to/recorded by RPNGC, e.g. Robbery, Robbery MV, Breaking commercial, breaking domestic, Stealing MV, stealing, Arson, Fraud. NB: robbery and robbery MV are counted twice- in both violent crimes and in property crimes. The extent and forms of violent crime vary by geographical area. Significant differences between rural and urban areas in the make-up of violent crime are found in the World Bank team’s analysis of the RPNGC data (see pie charts 7 and 8). In the Western Highlands Province (WHP) (Chart 7), rape and serious sexual assault account for the majority of violent crimes. In NCD, robbery accounted for nearly half of all violent crimes. In WHP robbery was only the third most common crime reported, with the same number of incidences of murder reported as there were for robbery. Again, the limitations of the data set need to be considered in interpreting this pattern. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 13 Pie chart 7: Western Highlands, reported incidences of violent crimes, 2010 1 23 51 murder + manslaughter 74 rape and other sexual offenses robbery + robbery MV GBH 117 firearms 51 abduction Source: RPNGC data for NCD and Western Highlands Province, 2010. Pie chart 8: NCD, reported incidences of violent crimes, 2010 30 36 125 murder + manslaughter 398 323 rape and other sexual offenses robbery + robbery MV GBH firearms abduction 855 Source: RPNGC data for NCD and Western Highlands Province, 2010. There is little data that can help determine the prevalence and the geographical spread of ‘ethnic-fighting’. The RPNGC data doesn’t allow for easy interpretation as to whether and where fighting between clans takes place. It is possible that large year to year fluctuations in assault and homicide rates in parts of the country- in the Western Highlands for example- are linked to clan fighting. Data from two separate analyses of hospital records in the Southern Highlands Province, reviewed for two different time- periods, found that approximately 25% of admissions for injuries inpatients may have 14 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea been sustained during clan-fights (Winnington 2008, Mathew 1996) 17 . Smaller scale studies and general commentary have indicated that clan fighting continues to take place in rural areas, and also in towns and cities (see Paper no. 2 in this series). Our research found that ethnic fighting does takes place in urban areas, particularly in settlements with mixing of migrants from different ethnic groups around the country. Disputes between rival ethnic groups may be played out between wantoks in either rural or urban areas, and transfers easily between geographical locales, regardless of where the dispute was initiated. Land issues in urban areas seem to be a significant trigger for violence on a large scale. VIII. The use of firearms in violent crime is increasing The evidence reviewed and the consultations undertaken for this brief does not provide a clear picture on the prevalence nor on trends, but does suggest that the use of firearms may be increasing, is intimately linked to the rise in violent crime and is not confined to the major urban areas and SHP. The availability of firearms, it is Much of the available literature on crime and violence in PNG points to the central role of firearms in the perpetration of crime and violence (see for example Muggah 2004, Haley and Muggah 2006, Haley and May 2007, UNHABITAT 2004, Walters 2007, Goldman 2007). Haley and Muggah (2006) reported that weapons were used in 80% of all violent crimes in NCD and SHP in 2006, and noted that while the number of guns in circulation compared with the population is low, firearms have become a significant factor in the increase in violent crime in PNG. Studies have also linked the use of arms to gender-based violence (Haley 2005: 35; Haley & Muggah 2006: 170, 174). Haley and Muggah (2006) noted that In the six months prior to the survey, over 8% of the respondents in SHP and 3% of respondents in NCD reported that someone in their household had been the victim of sexual assault or rape, which often involved the use of a weapon, such as firearms and bush knives Firearms may be acquired by the public in a number of ways. Anecdotal evidence suggests an increasing proportion of firearms are factory-manufactured rather than home-made. A porous border with Indonesia and poor chain of custody compliance and enforcement of firearms by police and security agencies have been pointed as ways in which firearms are transferred to the civilian population. Consultations conducted by the World Bank team also revealed that experts believed that significant numbers of 17Recording of information by health clinics however is complicated by the a fighting fee imposed on those seeking medical attention for injuries sustained through inter-clan disputes, and by the fact that medical attention is unlikely to be sought for most injuries, even those serious in nature (Kopi et al 2010). Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 15 firearms had been brought into the country by political candidates prior to the 2007 elections, and remain in circulation. According to the RPNGC crime statistics, firearms offenses are on the rise since 2008 (see Graph 6). They are most prevalent in urban, compared with rural, areas, with approximately 90% of all the reported firearms incidences in PNG occurring in NCD, and 7% occurring in the Highlands. Graph 6: Number of firearms offenses reported nation-wide is increasing since 2008 200 No of firearms offenses 150 100 50 0 2004 2005 2006 2007 2008 2009 2010 Year Source: RPNGC crime summary statistics. There are indications from a number of small- studies and analytical work that the use of firearms in crime in PNG could be under-estimated by the existing data. An analysis of the official Daily Crime Summary Narratives for 2010 18 however finds that the way in which crime reporting is recorded does not capture the majority of offenses where firearms are used, so the actual crimes occurring with the use of firearms may be far higher. For 2010 for example, analysis of the Daily National Crime Summary Narratives found that 82% of the reported crimes involving the use of firearms were not captured in the statistical summary of crimes reported 19 . The urban victimization surveys tell a similar story. The NCD survey for 2010 found that 9% of households surveyed had experienced crime that involved the use of firearms, higher than the 2008 figure of 5%. Other urban areas are also similarly affected. Kainantu, a small town in the 18PM&NEC and UNDP (2010) 19The daily crime summary narratives are completed by the provinces and then centrally collated and aggregated into national crime summary statistics. The PM&NEC and UNDP report that not only is there a loss of valuable crime information in the aggregated data, but data integrity is likely to be affected by capacity constraints of those compiling the data (PM&NEC and UNDP 2010). 16 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Eastern Highlands, was identified by the crime victimization studies as experiencing the highest rates of firearms violence of all urban areas during 2004-2008 (JAG 2008.) The most recent Lae crime victimization survey also notes a large increase in the use of firearms in 2010 compared with 2008 (LJSS, Lae, 2010). The only reduction in the use of firearms in the crime victimization surveys was reported for Kopoko. A survey of young people and crime in POM (UNHABITAT 2004) found that more than 60% of young people involved in burglary reported using a weapon in conducting the crime, but the type of weapon was not specified. Similarly, the Lae 2010 victimization study found that weapons -such as knives, iron bars, sticks and guns- were used in 45% of victimization cases in 2010 (LJSS Lae 2010). Data on injuries treated by the SHP provincial capital, Mendi, reports that a high number of injuries are gunshot wounds (Winnington 2008). IX. Little is known about perpetrators of crime and violence, other than that they tend to be male, and known to their victims Whilst the RPNG crime summary statistics do not provide information on perpetrators of the crime reported, the crime victimization studies do yield some important information on alleged perpetrators of the crime victimization reported in the surveys. The most recent studies for Lae, Kopoko and NCD (LJSS Lae 2010, LJSS Kopoko 2010 and LJSS NCD 2009) all find that in the majority of ‘the most troubling cases’, victims reported that they knew the perpetrator. The perpetrators were most likely to be ‘wantoks or relatives’ or neighbors. More than one perpetrator was also reported in the majority of crimes reported in the victimization surveys 20. This is likely to be a factor in low reporting rates, as traditional methods of dispute resolution may be preferred in crimes where the victim and perpetrator know each other. Young perpetrators of crime, as found by a survey of young people in POM (UNHABITAT 2004), are more likely to be male (75%, compared with 25% female). Women are involved more intermittently, or infrequently in crime, and are more likely to be engaged in spontaneous, petty crime, as individuals, rather than in a group. Men are more likely to be regular, repeat offenders 21. The study provides some interesting data also on gender differences in life ‘coping-strategies’ employed by young men and 20 For example, for 69% of crime victimization reported for Lae in 2010, there was more than one perpetrator. In NCD in 2009, the figure was 67%. 21 The UNHABITAT survey of youth is one of the studies that looks at crime perpetration as well as victimization. The findings do not provide any data however on the prevalence of crimes committed by young people. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 17 women, noting that young men did not use the same kinds of networks and outlets for the stresses of life as do young women. The UNHABITAT survey also found that young male offenders in POM often commit crimes as a gang, with preparation, using force or violence and often whilst using alcohol or other substances. Many of the other qualitative studies, particularly those on gender-based violence also link the use of alcohol and drugs to the initiation of violence and the perpetration of crimes (this is discussed in more detail in Briefing Note 2). X. Arrest rates have remained consistent and low, even in years when reported crimes have decreased The RPNGC data reveals that while the total number of crimes reported has experienced a general, slow, decrease, arrest rates 22 have stayed low, and as a percentage of total crimes reported have not changed significantly. In PNG, arrest rates have stayed lower than 50% of reported crimes for the entire period. The data appears to show a fairly consistent correlation between total crimes committed and the percentage of arrests 23. Such consistency between crime rates and arrest rates is unusual, particularly that arrest rates stay low even when crime rates decrease. In discerning the reasons for this, it could be important to look, for example, at police financial and human resources over the same time period, and at rates of confidence and trust on the part of communities. Understanding the consistency of the relationship between crime and arrests, regardless of the total numbers of crimes is important. Our consultations with the RPNGC and other stakeholders found that one factor that contributes, at least in part, to the trend- is that of a form of ‘forum shopping’ or ‘forum leveraging’ by justice seekers. Victims of crimes in some cases will file an incident report with the police—triggering the recording of a crime, often with the intent of simply using the incident report as leverage in negotiations to resolve the conflict through other means. In other case, victims may simply change their mind, or have the conflict resolved through other means while police investigation is being undertaken. This use of alternative forums however is likely 22 Arrest rates often provide a useful measure of law enforcement. However, they should be considered with caution, as they can be misleading. For example, the arrests recorded for a particular category of crime do not necessarily correspond with the crimes reported within that same time period as arrests often take place a considerable amount of time after a crime is reported. Multiple persons can be arrested for the same crime, and sometimes reported crimes can be addressed without any arrests being made. 23 For example, arrests dropped from 40.5% of total crimes in 2003 to 27.5% of total crimes in 2004, in the same year that the total number of crimes dropped by more than 40%. For 2008 to 2010, arrest rates increase slightly year on year. 18 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea only a partial explanation of the trend between arrest and crime rates. The large average distances between police stations is also likely to be a significant factor in both victims reporting crimes, and in police investigating crimes once reported. Further analysis on the changing nature of crime noted in the section above (the increase in violent crime as a proportion of total crime) to see if this has a bearing on the arrest rates could be another relevant inquiry to explore in this regard. The indication is that the low arrest rate for violent crime is a factor in low arrest rates overall 24. Graph 7: Comparison of crime reported and arrests made (numbers), 2000-2010 14,000 Total no. of crimes reported Crime reported and arrests made 12,000 Total arrests 10,000 (number) 8,000 6,000 4,000 2,000 - 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year Source: RPNGC crime summary statistics. The RPNGC data shows that, for certain crimes, arrest rates are consistently far lower than for other crimes. Arrest rates for Grievous Bodily Harm (GBH), rape, other sexual offenses, robbery, and breaking and entering have been generally low. Given that robbery, rape and GBH are amongst the most commonly reported crimes, the particularly low arrest rates warrant further analysis. A clearer understanding of the victims, perpetrators and intentions, as well police responses may be informative. Qualitative data collected through interviews with victims who have reported these offences, analysis of RPNGC crime narratives and interviews with case-assigned police officers themselves would help interpret this trend. 24The total number of crimes and arrest rates have been calculated removing drugs and firearms offenses, as the majority of these offenses are detected and recorded only when an arrest is made. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 19 Graph 8: Reported crimes and arrests compared 1,400 1,200 1,000 800 No. crimes reported 600 No. of arrests 400 200 0 GBH Rape Murder Robbery Source: RPNGC crime summary statistics, 2010. Arrest rates however, only tell part of the story of law enforcement. By themselves they are not indicators of the efficacy of the justice system in addressing crime and violence. Arrest rates would need to be cross-referenced with data on cases prosecuted through the courts in order to determine whether crime is being effectively addressed through the formal justice system. Upon preliminary analysis we can see that cases brought to the formal court system 25 are far lower than the, already low, total number of arrests made during a given year. For example, the number of arrests made in the first half of 2011 in Lae totaled 654, while during 2011 only 69 cases had been brought to the formal court system, four of which were crimes committed during the end of the previous year. Our consultations also found that the hybrid village courts, despite their limited jurisdiction, were commonly used to resolve criminal as well as civil matters. Victims would seek resolution through the village courts even when the crime had been reported to the police. XI. Conclusion Summary of findings There is clear evidence from multiple sources that overall levels of crime and violence in PNG are high compared with rates found elsewhere in the World, but that crime and Data on cases passing through the formal Court system can be obtained from National Court 25 Circuit Lists. The data noted above is from the Lae National Court Circuit list. 20 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea violence may be stabilizing. However, certain areas of the country have far higher crime rates than others. While data for rural areas is thin beyond that recorded by the police, the data that is available suggests that there is a concentration of crime in urban areas- both large and small, although crime and violence are by no means confined to urban areas. There is also evidence that some of the higher-crime areas have not experienced the same decrease in crime rates as the national average. Furthermore, analysis of the RPNGC data finds that the high-crime areas have seen an increase in violence used in the perpetration of crime. These forms of violence, preliminary analysis suggests, vary from place to place, with property crimes committed using violence more prevalent in urban areas, and inter-personal violence more common in rural areas. While vulnerability to crime is high for men, women are even more likely to be victims of crime, in particular gender-based crimes committed by perpetrators known to them. Under-reporting of gender-based violence, as well as for crimes more broadly, appears to be chronic. Formal law enforcement responses to high crime in PNG appear to be weak and effected by forum-shopping, resulting in consistently low arrest rates, and low rates of prosecution in the courts. Gaps in data and knowledge • Much of the information on crime trends comes from a limited body of literature that, in general, does not refer to robust empirical evidence; • There is a bias in the primary research studies towards major urban areas and the Highlands, with a dearth of studies conducted outside these areas. At the same time, the official RPNGC data and our consultations indicated that there is likely to be great variation in the incidences and experiences of crime and violence across the country. • The regularly-conducted crime victimization studies do not exist for all urban areas, and are not conducted at all in rural areas of PNG. As more than 80% of the population lives in rural areas, this represents a significant gap in data availability. • The two main longitudinal data sets- crime statistics from RPNGC and data from the urban crime victimization studies- show very large divergence in the numbers, and to some extent, the types of crimes occurring. This complicates the task of arriving at realistic estimates of crime and violence. Significant methodological, data quality and data integrity issues are likely to exist with some of the data sets reviewed, including the RPNGC data. Methodological challenges are also likely to compromise the accuracy of some aspects of the crime victimization data. In particular, these two data sets are unlikely to have been able to capture the actual prevalence of gender-based violence. Review and refinement of the survey instrument for the crime victimization surveys and for the data collated for RPNGC national crime summary statistics would help address these data issues. • The incidence of tribal or ‘ethnic-fighting’ is not captured in the RPNGC data, nor in the victimization surveys, and as such, it is difficult to ascertain beyond local Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 21 reporting, hospital records and the media to what extent this form of violence continues in both rural and in urban areas. • Commentary on crime and violence can become quickly outdated or inaccurate due to changes in crime rates, and seemingly random fluctuations within urban areas; • Varying levels of crime and differences in crime profiles across the country make it difficult to make generalizations about crime nationwide; • The relationship between crimes committed and arrests made is little understood. The fact that arrest rates have remained low even during years when reported crime rates have decreased, deserves further analysis. Potential factors such as the fluctuating resources invested in policing and criminal justice services and administrative changes in police deployment and data recording, confidence in the police and the availability and use of alternative fora, and other political and social dynamics have on this relationship. • With a large percentage of all crime being violent crime -for which arrest rates are lower than for other crimes, this may be a significant factor in low arrest rates overall. Works cited Barker, P. Address to ONG Institute of Directors, Workshop on ‘The Fight Against Crime’, 27th August 2009, To Robert Training Centre. Bradley, C. (2001) Family and Sexual Violence in PNG: An Integrated Long-term Strategy, Discussion Paper No.84, Institute of National Affairs, Port Moresby. Dinnen, S., D. Porter and C. Sage (2011) Conflict in Melanesia: themes and lessons, World Development Report 2011, Background Case Study. Washington: World Bank. Eves, R. (2006) Exploring the Role of Men and Masculinities PNG: How to address violence in ways that empower both men and women, Caritas Australia. Eves, R. (2010) Masculinity Matters: Men, Gender-Based Violence and the AIDS Epidemic in Papua New Guinea, in Vicki Luker and Sinclair Dinnen (ed.), Civic Insecurity: Law, Order and HIV in Papua New Guinea, ANU ePress, Australia, pp. 47-79 Goldman, L. (2007) “Ho-ha in Huli’: Considerations on commotion and community in the Southern Highlands, in Haley, N and R. May (eds) Conflict and Resource development in the Southern Highlands of Papua New Guinea in Conflict and Resource Development in the Southern Highlands of Papua New Guinea, State society and Governance in Melanesia Program, Studies in State and Society in the Pacific, No. 3, Canberra: ANU. 22 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Haley, N and R. May (2007) (eds) Introduction: Roots of conflict in the Southern Highlands in Haley and May (eds) Conflict and Resource development in the Southern Highlands of Papua New Guinea in Conflict and Resource Development in the Southern Highlands of Papua New Guinea, State society and Governance in Melanesia Program, Studies in State and Society in the Pacific, No. 3, Canberra: ANU. Haley, N and R. Muggah (2006), Jumping the Gun: Armed Violence in Papua New Guinea, in Small Arms Survey 2006: Unfinished Business, Geneva: Small Arms Survey. JAG (Justice Advisory Group) (2008) Urban Crime Victimization in Papua New Guinea, 2004-2008: A synthesis, 17 November 2008, PNG: Justice Advisory Group. Lewis,I., B. Maruia, D. Mills, D. and S. Walker (2007) Final Report on Links Between Violence Against Women and the Transmission of HIV in PNG, University of Canberra. Lewis, N. (2007) Conflict Vulnerability Assessment of the Southern Highlands Province, in in Haley and May (eds) Conflict and Resource development in the Southern Highlands of Papua New Guinea in Conflict and Resource Development in the Southern Highlands of Papua New Guinea, State society and Governance in Melanesia Program, Studies in State and Society in the Pacific, No. 3, Canberra: ANU Mathew, P.K. (1996) ‘Changing trends in tribal fights in the highlands of Papua New Guinea: a five year review’ in Alpers M. (ed), Papua New Guinea Medical Journal, Volume 39, No 2: 117-120. Medicines Sans Frontiers, No date, Hidden and Neglected: the medical and emotional needs of survivors of family and sexual violence in Papua New Guinea, MSF- PNG: Port Moresby. Muggah, R. (2004) Diagnosing Demand: assessing the motivations and means for firearms acquisition in the Soloman Islands and Papua New Guinea, State society and Governance in Melanesia, Discussion Paper, 2004/7 ANU LJSS (2010) Kokopo Urban Community Crime Victimization Survey, 2010, Law and Justice Sector Secretariat, Monitoring and Evaluation Unit. LJSS (2010) Lae Urban Community Crime Victimization Survey, 2010, Law and Justice Sector Secretariat, Monitoring and Evaluation Unit. LJSS (2008) National Capital District, Urban Community Crime Victimization Survey, 2008, Law and Justice Sector Secretariat, Monitoring and Evaluation Unit. LJSS (2009) National Capital District, Urban Community Crime Victimization Survey, 2009, Law and Justice Sector Secretariat, Monitoring and Evaluation Unit. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 23 PM&NEC and UNDP (2010), Making sense out of data: An assessment of daily crime summary in PNG. Joint project of the Department of the Prime Minister & National Executive Council (PM&NEC) and the United Nations Development Programme (UNDP). Port Moresby. Seibert, S. (2009) Ending Family and Sexual Violence in Papua New Guinea: A review of the National Strategy and Recommendations, UN (2010) Report of the United Nations Special Rapporteur on Torture and other Cruel, Inhuman or Degrading treatment or Punishment to the UN General Assembly, Mission to Papua New Guinea, 7 February 2010, A/hrc/16/52/add.5 UNDSS (2005) National Crime Summary Report, 1 Jan 2004-31 Dec 2004, Port Moresby: UNDSS UNHABITAT (2004) Youth and Crime Survey, Port Moresby, Papua New Guinea, 2004 UNODC (2011) Global Study on Homicide, Vienna: UNODC. UNODC and UNECE (2010) Manual on Victimization Surveys, Geneva: United Nations WHO (2002) World Report on Violence and Health, Geneva: WHO Winnington, A. (2008) Small Arms Violence in the Highlands of Papua New Guinea: Analysis of Hospital Related Data 1999-2007, University of Otago, Wellington, New Zealand. World Bank (2010) Violence in the City: Understanding and Supporting Community Responses to Urban Violence. Washington DC. 24 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Annex 1: Data on crime and violence 1. Data sources and gaps The two largest sets of data on crimes reported and victimization rates for PNG are the official statistics from the Royal PNG Constabulary (RPNGC), covering the entire country, and unofficial victimization surveys for major urban areas, conducted by the Law and Justice Sector Secretariat. A third, nationwide data set, derived from the 2009 Household Income and Expenditure Survey is expected to become available in the coming months. Other data sources include the National Courts, academic literature, and both quantitative, qualitative and mixed-methods studies undertaken by international organizations and NGOs. The two, currently largest, data sets are highly likely to suffer from differing forms and degrees of under-reporting, as well as issues of data integrity with the RPNGC statistics. The smaller studies such as Eves (2006) and Haley and Muggah (2006) utilize strong methods of data collection and triangulation, but are limited to certain geographical areas. Very little data is available on crime and violence outside of the main urban areas and the Highlands. Further, knowledge of urban crime and violence outside of the NCD is based largely on the crime victimization surveys. The main data sources are outlined below. Official crime data from the RPNGC records the type and number of crimes reported, and the number of associated arrests. Data is available for crimes committed from 2000 until the first half of 2011, as both as raw data and classified by crime category, provinces and regions. The RPNGC data provides rich data on the prevalence of each crime over time and spatially. The data on numbers of arrests can be used to establish arrest rates over time and for different types of crime. However, the data suffers from gaps in recording- many provinces have at least one category of crime for which no data is recorded. Aggregation of the daily crime summaries from each province has seen a loss of data quality and the data is not gender disaggregated. Analysis of daily crime summaries is provided for 2009 and 2010 under a joint initiative by the Department of Prime Minister & National Executive Council (PM&NEC) and UNDP in order to establish a baseline on firearms offenses as reported in the RPNGC data. The reports find that even the vast majority of firearms offenses reported are not captured in the RPNGC crime statistics. Official data from the National Courts on cases handled by each court during a given time period. Date committed, the charge, address of the defendant and other remarks are all available for each defendant. The data fields are limited, and as such doesn’t provide much depth for analyzing crime arrest and prosecution trends. Official data from health service providers is analyzed in one study in particular which looks at the injuries treated in the hospital in Mendi, provincial capital of the Southern Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 25 Highlands province (Winnington 2008). The study provides data on intake, severity of injury, cause of injury (intent, including whether the cause was inter or intra-tribal fighting) and cost associated with the injuries. This data is gender disaggregated. A Household Income and Expenditure Survey for the entire country was conducted in 2009. The surveys included a series of questions on disputes (type, who with, damage or impacts as a result of the dispute) and dispute resolution. Questions on health expenditures, including for traditional practitioners and herbs, and theft, funeral expenses, amongst others, could yield some useful data in looking at crime and violence. Only preliminary findings of the HEIS survey were available at the time of writing this note. Victimization surveys for major urban areas including Lae, National Capital District (NCD), Kopoko, Mt. Hagen, Kainantu, Goroko, Arawa and Buka. Commencing in 2004, 19 such studies have been undertaken to date by the Monitoring and Evaluation Unit of the Law and Justice Sector Secretariat. They involve a fairly large sample 26 and provide statistical data on community perceptions about crime, and the levels, extent and type of actual crime victimization. As the studies for the largest urban areas were conducted more than once, data for these towns and cities is comparable over time. The data from the victimization studies are also valuable in comparing victimizations rates with the rates of crime reported, as such establishing the degree of under-reporting. Direct comparison is not possible however for most categories of crime as they differ significantly from the crime categories used by RPNGC. The only exception is that of ‘killings’ as found by the crime victimization surveys and ‘murder’ and ‘manslaughter’ reported by RPNGC. The data from victimization studies may be able to provide information on what kinds of crime are less likely to be reported to the police and why. Socio-economic factors related to crime victimization are also provided in these data sets. Gender disaggregation and information from the perspective of the victims and about the perpetrators significantly furthers our understanding of crime and violence. Business Victimization studies. Two victimization studies have looked specifically at the prevalence and impacts of crime and violence on businesses in Port Moresby. The first was conducted in 2005, and the latest in 2010. The WB team will be conducting their own study on the impacts of crime and violence on business in PNG in 2013. The Prison Population Briefing for PNG. The Prison Population Briefing, prepared by an NGO of the same name, provides summary information on the number of persons in correctional facilities, holding cells, the capacity of these facilities are filled with prisoners, the number of juveniles in detention and the number of pre-trial detainees being held. 26For example, the 2009 NCD study surveyed 593 people aged 15 years and over in 265 households, located in eight diverse sample sites across NCD. 26 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea A survey on youth and crime in Port Moresby in 2004 27 includes perceptions of crime, crime incidence and typology, the perpetrators’ profile, the use of violence and weapons in criminal activities and the use of alcohol and drugs in criminal activity. The UNHABITAT study presents findings on the link between crime and unemployment, crime and substance-abuse, as well as presenting some figures for rates of criminal activity amongst young people (i.e. the percentage and types of crimes committed by young people compared with other groups). It also provides gender disaggregated data for most of the issues covered. While the UNHABITAT study looks at rates of criminal activity that involve young people, motivations for engagement in criminal activity are not asked directly of respondents, but are rather inferred from other data collected. Other studies and literature: There are other studies which provide useful qualitative, and some, often limited quantitative background material. The Pacific Youth literature review conducted for the World Bank in 2009 looks at youth engagement in risky behavior and crime, as well as presenting relevant demographic information. The Small Arms Survey (Haley and Muggah 2006) looks at the use of firearms in NCD and the southern Highlands province, through victim-oriented sample studies employing a triangulation method. There are a number of studies which report on domestic and gender-based violence, using varying degrees of quantitative data, but with rich qualitative analysis (such as Lewis et al 2007, Eves 2006 and 2010, Seibert 2009, UN Women, No date, and Bradley, No date) and on the use of weapons in this context ( MSF, Haley and Muggah, and Lewis 2007). Academic and general commentary from anthropological papers and policy documents was also drawn upon for some parts of the desk-study. 2. Challenges in classification of crime What is considered to constitute a crime varies from society to society. Crime classification and categories, both in official data and from victimization studies can tell us what is considered a crime, and as such reflect the societal norms and values around certain behaviors. In PNG, for example, there is currently no legislation that criminalizes domestic violence 28 and a separate category for domestic violence doesn’t exist within the RPNGC data on crimes reported. The categories used by the RPNG to record crimes reported have changed over time. Notably, rape and manslaughter were added as categories in 2004, along with those of arson, bribery, and abduction data. Similarly, conceptualizations of crime and violence, 27 The study, undertaken by UNHABITAT, used a targeted actor-oriented approach, looking at 1,500 young people aged from 15-35 years old. 28 Cases of domestic violence fall under the provisions for common and aggravated assault under the Criminal Code, however, there is a ‘Standing Order ’to the RPNG constabulary, to treat with seriousness incidences of domestic violence. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 27 and as such, categorization, vary from study to study 29. In general, however, most of the studies reviewed in this paper follow a similar way of classifying crimes. There are a few important exceptions however, which may present challenges in the comparison, analysis and costing of crime and violence due to overlapping classifications and multiple meanings employed for crime terminologies. The two largest sets of data on crime and victimization rates- one official from the Royal PNG Constabulary covering the entire country, and the other unofficial victimization surveys for Lae, NDC and Kopoko have some significant differences in the categories they use for crimes 30, making difficult a direct comparison of crimes reported to the police with those reported by respondents in the crime victimization studies. Another challenge that presents itself is the overlapping of different forms of crime and violence. Though burglary, breaking and entering and robbery are considered property crimes, the victimization studies found that violence was often used in conducting these crimes. The UNHABITAT survey on youth and crime, for example, uses the term ‘burglary’ for crimes involving entering a property and stealing from it. The survey notes however, that in Port Moresby most burglaries involve violence or the threat of violence 31. The urban victimization studies recognize the existence of multiple forms of violence or crimes in a single act, and as such count them separately. So for example, a burglary conducted using violence would have been recorded in the victimization studies as two incidences, one being the burglary, the other being unprovoked violence. This method of recording may significantly inflate the victimization rates estimated by the study, compared with official data from the RPGNC. 29 One interesting example is the use of the category ‘provoked violence’ by the JAG urban crime victimization studies, indicating that differentiation between crimes associated with payback, or cycles of retributive violence, and those ‘unprovoked’ acts of violence, is important to respondents. 30 The urban crime victimization studies used 12 categories, while the RPNG data was sorted according to 19 categories, from 2004 onwards. The victimization surveys included Property stolen, Breaking and stealing, Assault, Destruction to property, Firearm use, Vehicle theft, Stealing with force, Unprovoked Violence, Provoked Violence, Sexual assault, Killings, and Other. In addition, after 2006, the victimization studies included data on children as victims of crime. The RPNG data categories are as follows: Stealing of Motor Vehicle, Stealing >K1000, Breaking and entering commercial property, Breaking and entering domestic property, Grievous Bodily Harm, Robbery, Robbery of a motor Vehicle, Arson, Firearm Offenses, Abduction, Rape, Other sex offenses, Murder, Manslaughter, Drug offenses, Bribery, Fraud, Escape from prison/CIS, and Escape from police. 31 Of the young respondents in the UNHABITAT survey (2004) who had been arrested for burglary, 85% admitted the use of weapons and 66% notes that they actually used violence. Similarly, of the respondents who have committed burglary, but were not arrested, 88% claimed to have used weapons and 63% note they actually used violence. Eves notes that over 40% of recorded crime involves violence. 28 ■ The Socio-economic Costs of Crime and Violence in Papua New Guinea Particular challenges exist in the conceptualization of gender-based violence. “Domestic violence,” “gender-based violence,” and “family and sexual violence” are some of the terms used by researchers and policy-makers in the studies reviewed. While “family and sexual violence”, as used in reports by Medicins Sans Frontiers (MSF, no date), Seibert (2009) and UNHABITAT (2004), predominantly refer to crimes which victimize women, it also includes those crimes against young men under 15 years and children . In fact the UNHABITAT survey found that nearly half of all young sexual abuse victims are male. The danger is that while gender-based violence may encompass all forms of crimes and victims (either male of female) this term is used interchangeably with ‘violence against women’ and as such excludes from analysis, costing and intervention, an important social group subjected to violence in PNG- that is young men under the age of 15years. Another challenge is the inclusion of prostitution as a crime in the UNHABITAT survey. While based on the criminal code, it runs the risk of conceptualizing as crimes, actions which, other commentators have considered as a form of victimization of gender-based crimes (UN Women, Lewis et al 2007). State-sanctioned violence 32 or aggression and transgressions by the state security apparatus is considered to be a problem, particularly in areas around mining and logging industries, however, there is little hard data-based evidence both as to its prevalence, and even less to its impacts and costs. Use of violence by private security forces-particularly those employed by resource companies- against local communities , has also been highlighted as a form of violence condoned by the state. These forms of state sanctioned violence are neither explicitly acknowledged nor covered by the crime surveys other than that on Youth and Crime. The survey includes data on ‘physical abuse’ by the police towards young offenders in police custody (UNHABITAT 2006), however the study does not ask any further questions nor discuss possible impacts. Given the high rates of detention of prisons, the ill-treatment of prisoners 33 is a relevant and perhaps significant form of state-sanctioned violence. 32 The broader definition of ‘State-sanctioned violence’ includes acts of violence not only undertaken by state apparatus, but also by private actors against whom the state is unwilling to take reasonable action in the spheres of law, policy and sanctions. In the context of PNG state- sanctioned violence might also include the use of the Death Penalty and domestic violence (Cf. UN 2010). 33 The report of the UN Special Rapporteur on torture and other cruel, inhuman or degrading treatment or punishment, 2010, notes extended detention, severe beatings, overall lack of medical attention and poor sanitary conditions for prisoners in police detention and in correctional facilities, in some cases, constituted state sanctioned torture and ill-treatment. Paper 1: Trends in Crime and Violence in Papua New Guinea ■ 29 30