88599 Research and Dialogue Series Note No. 3 Crimes and Disputes: Missed opportunities and insights from a national data collection effort in Papua New Guinea World Bank LEGJR, Legal Vice-Presidency February 2014 © 2014 The World Bank Group 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org/justiceforthepoor E-mail: j4p@worldbank.org All rights reserved. This note, the 3rd in a series, 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 Papua New Guinea. It was drafted by Sakuntala Akmeemana, with significant inputs by Reno Diwa, Nicholas Menzies and Laura Bailey. Timothy Bulman, Alys Willman, Sadaf Lakhani and Amarachi Utah made useful comments and contributions at various stages of the process. This work was requested by the Prime Minister and is being undertaken with extensive input from international partners and local stakeholders. The larger study team is comprised of Alys Willman, Sadaf Lakhani, Gary Milante, Nicholas Menzies and Sakuntala Akmeemana. This paper has not undergone the review accorded to official World Bank publications. 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. The World Bank does not guarantee the accuracy of the data included in this work. 2 I. Introduction As in many developing countries, data collection has proved to be a considerable challenge in Papua New Guinea (PNG). The country has contended with what are said to be amongst the highest levels of crime and violence in the world, and there have been ongoing Government and donor interventions in the law and justice field for several decades. Yet, for the most part, these efforts have been undertaken without a strong evidence base and without reliable data on crime rates. They lack a clear understanding of the most frequent or serious legal disputes faced by citizens, or indeed which institutions are actually used to obtain redress for the legal wrongs they experience. Where interventions do not have a strong empirical base, they are difficult to design for measurable success. A welcome effort at data collection on dispute incidence and personal security was made in PNG’s Household Income and Expenditure Survey (2009/2010) (HIES or “the Survey�),1 an experience that also highlights some of the challenges of such an exercise. For the first time, the HIES asked questions about dispute incidence and personal security. Along with Timor-Leste, which included a disputes module in an extension of a living standards survey in 2007, PNG was one of the first countries globally to include such questions in a HIES, as opposed to a dedicated survey on justice needs and redress. The fact that the two survey modules on disputes and personal security were part of a larger household survey has the benefit of allowing correlations between demographic and socio-economic data with those on dispute and crime incidence, security and access to redress. However, being structured as part of a larger survey possible limited the number of questions that could be asked on disputes and crime, and in turn limits the depth of the analysis. The Survey was administered in all provinces of PNG to a nationally and regionally representative sample of over 4,000 households,2 and provides comprehensive data about the socio-economic status of households. The results are representative at the level of the country’s four regions (Southern, Highlands, Momase and Islands) plus the ‘Metropolitan Area’ (which comprises the major urban areas of National Capital District and Lae), as well as for a rural/urban breakdown within each region.3 The results are not representative at the provincial level. This Briefing Note will highlight some of the main findings on dispute and personal security, noting which results need to be treated with caution. It also distils a number of lessons from this data collection effort. As with all surveys, the questions need to be read carefully in interpreting responses (The questionnaires for each module can be accessed at http://go.worldbank.org/IN6EW8T9P0). There are some important differences in who was asked the questions in the two modules and some overlap in 1 The 2009/10 HIES is the first undertaken in PNG since 1996. The National Statistics Office has released summary tables and the full results of the HIES are publicly available for interested scholars at http://catalog.ihsn.org/index.php/catalog/3243. A summary of the major findings can be found in the PNG Economic Briefing (World Bank 2013). 2 The original sample size was 4,191 households. After removing incomplete, missing, and improper results codes, there were 4,081 households in the sample. 3 Highlands region includes the provinces of Western Highlands, Southern Highlands, Eastern Highlands, Enga, Chimbu, Hela and Jiwaka; Islands region comprises East New Britain, Manus, New Ireland, Bougainville, West New Britain; Momase comprises West Sepik, Morobe, Madang and East Sepik; the Southern region comprises the National Capital District, Western, Gulf, Central, Milne Bay and Northern Oro. 3 the types of issues covered by them. This note does not report certain results where there is significant cause to believe they are implausible. Getting Better Dispute and Crime Data in Future Household Surveys The most robust results from these PNG HIES data relate to questions about the fear of crime. Both crime and dispute incidence figures, as well as the section on dispute resolution mechanisms need to be treated with some caution due to problems in the questionnaire design: - Many of the questions in the personal security module were not framed in accordance with domestically or internationally recognized categories of crime, and omitted important categories. In particular, this limits the comparability of the results on crime incidence with international data or other domestic data sources. - The categories in the disputes module were clearly distinct, with considerable overlap amongst them and thus potential for confusion amongst respondents and inconsistent categorization of similar disputes. - The disputes module was adopted from a questionnaire used for a very specific purpose: a baseline survey for a community paralegal program in Indonesia. As such, it was not intended to be used to infer rates of dispute and crime prevalence for a national population in another country context. - The questionnaire was similarly not intended to explore levels of satisfaction with various dispute resolution mechanisms in the PNG context, where over 90% of cases remained unresolved at the time of the survey. In the event of an unresolved dispute, the questionnaire asked no further questions about the resolution of the dispute or the respondent’s levels of satisfaction with the forum that was used, which severely limits the ability to further interrogate this data. Questionnaire design should commence with a clear conceptual or analytical framework which highlights the main analytical questions that are sought to be probed and answered through the survey. While survey instruments from other countries can provide good points of reference in relation to framing certain questions, in particular to allow for a point of comparison with international data, they cannot provide the underlying logic for the survey design. Customization for country context should not detract from the importance of using nationally and internationally recognized standard categories of crimes, in order to preserve the ability to make robust comparisons using the survey data. A good questionnaire and a well-executed survey can provide a broad outline of the picture of dispute and crime incidence and use of dispute resolution institutions. There is a wealth of further information and insights that remain to be gained from qualitative research and analysis. 4 Key findings  Disputes are experienced more frequently by households in rural than urban areas, with the Momase and Highlands regions having the highest rates and the Islands region the lowest.  Forty percent of households experienced a dispute in the year before the survey, with the main sources of disputes relating to land, water and natural resources more generally.  Intra-familial disputes and violence are reported more frequently by women and as having a greater impact on them. Men report disputes that relate to the public and economic sphere more often than women and as having greater impact upon them.  Disputes which lead to death are most often linked to land and ‘other tribal’ issues and are most common in the Highlands region. ‘Other tribal disputes’ are the most lethal category of disputes.  While over half (56 percent) of all respondents seek assistance in resolving the dispute in the last 12 months with the greatest impact upon them, over 90 percent of those disputes remained unresolved at the time of the survey.  Community leaders and village courts are most commonly used for dealing with disputes, with churches and NGOs relied upon relatively infrequently. The wealthiest quartile of respondents is least likely to seek the help of their wantoks*, and more likely to use village courts.  The fear of crime significantly affects freedom of movement, especially for women (55 percent, compared to 30 percent of men) and urban residents (59 percent, when compared to 40 percent of those in rural areas). While just over half the respondents reported that crime does not stop them from “doing anything�, there are large differences between men (63 percent) and women (40 percent) and respondents in urban (54 percent) as opposed to rural areas (32 percent).  Respondents are most afraid of drug and alcohol-related crimes (presumably because of the unpredictable violence associated with them) although personal property crimes appear to be most prevalent.  While household surveys inherently under-report the incidence of domestic violence, 9 percent of women report being ‘beaten’ by someone in their household in the last 30 days, and also report their husbands/partners being the perpetrator in two thirds of cases.  Roughly one in every eight crimes is reported to the police. * Wantok refers to the network of clan, family and tribal relations that are central in Melanesian cultures. 5 II. What were respondents asked about disputes and personal security? In the disputes module,4 the household head or his/her spouse was asked about the incidence of several categories of disputes in the previous 12 months, including those related to natural resources; development projects; government decision making; family, financial and criminal matters. The lay meaning of the word ‘dispute’ was used, including both crimes and civil disputes, but also capturing incidents that do not give rise to a legal remedy. In some instances, the distinctions between these categories are also unclear: for instance, most tribal disputes relate to land or bride price, so respondents may have had difficulties categorizing disputes. It is therefore likely that the residual category “other tribal disputes� captured some of these common dispute types. Respondents were asked about the identity of the other party to the dispute; whether that dispute resulted in property damage, injury or death; which of the disputes had the most severe impact upon the respondent or their household; whether help was sought to resolve the dispute with greatest impact; the nature of the assistance sought or dispute resolution mechanism used; and the relevant costs and levels of satisfaction with the dispute resolution process. The questions in the personal security module were posed to all individuals in the surveyed households above the age of 15 years. These questions relate to perceptions and fears about crime, the impact of fear on respondents’ daily lives and the incidence of certain crimes in the previous 12 months. Female respondents were also asked an additional set of questions about whether they had suffered physical violence in the home in the previous 30 days, the identity of the perpetrator and whether they had sought help. III. What did the HIES Disputes Module Show? What is the dispute profile in PNG? Forty percent of respondents reported that their household experienced a dispute (both crimes and civil wrongs) in the previous 12 months,5 with a family member being the most likely adversary (in 40 percent of cases). As in many developing countries, especially those that are predominantly agrarian, households are most likely to experience disputes related to land and water (29.8 percent).6 In fact, natural resource disputes (including land, water, forestry and agriculture) comprised 38.4 percent of all disputes. This is consistent with the qualitative literature about the intensification of struggles over access to, and control of, land and natural resources in the context of resource extraction in PNG.7 4 The questionnaires for each module can be accessed at http://go.worldbank.org/IN6EW8T9P0 (National Statistical Office 2009a & 2009b). 5 The PNG disputes module was based on that used in Timor-Leste in 2010 (Dale et al. 2010a), where 31% of respondents reported having experienced a dispute. 6 In Timor, 10% of households reported a land-related dispute in the prior 12 months (Dale et al. 2010b), whereas in a survey in Bangladesh, the figure was just over 20% of households in a two year period (Akmeemana 2011). 7 Allen (2013); Filer (2011). 6 Land disputes were followed by domestic violence (11.7 percent), theft (11.3 percent) and physical assault (9.5 percent). Men report disputes relating to the public and economic sphere more frequently than women, whilst the latter report intra-familial disputes and violence more frequently. Figure 1: National dispute profile: percent of disputes, by incident type Other tribal disputes Dispute over forestry Dispute over agricultural production Dispute over bride price Physical assault Theft Dispute over domestic violence Dispute over land /water 0 5 10 15 20 25 30 35 While the rates for domestic violence in the Metropolitan region (22 percent) surpass land disputes (19 percent), otherwise there appears to be little geographic variation within the dispute profile. Households are most likely to experience land and domestic violence disputes across the country, although more households in the Islands region appear to report land disputation (50 percent) than elsewhere in the country and lower levels of domestic violence (10 percent).8 Tribal disputes appear most prominently in the Highlands. Which disputes affect men and women most seriously? Respondents were also asked to identify the disputes that impacted them most seriously. As in many developing countries, where citizens are most seriously affected by adverse impacts on livelihoods, respondents indicated that land and water disputes have the most serious impact on their lives. Indeed, land and water disputes were cited by almost four times as many respondents (46 percent) as the next dispute category (domestic violence, at 12 percent). Disputes over natural resources had the most serious impact of the lives of male respondents in rural areas. Domestic violence was cited as having the greatest impact by almost twice as many female respondents (17 percent) as males (10 percent) and almost twice as many respondents in urban areas (20 percent) than rural areas (11 percent). 8 Compare the results of a recent United Nations report on violence against women, which suggest extremely high rates of violence in Bougainville (Fulu et al. 2013). 7 Where are disputes most prevalent? Figure 2: Disputation rate by region 60.0 A greater percentage of households in 50.0 rural areas experienced disputes (41 Dispute rate (%) percent) than in urban areas (29 40.0 percent), both in aggregate and across all 30.0 dispute types except for disputes with government and those linked with 20.0 financial matters and development 10.0 projects/aid flows.9 0.0 Momase region has the highest rates of Momase Highlands Southern Metro Islands disputation (with almost half the households experiencing a dispute) and the Islands the lowest. These results could be explained by the fact that the Momase region has significant economic activity (mining, agriculture, manufacturing and fisheries). Intense contests over land could be expected, and for these to manifest themselves in other types of disputes. At the same time, poverty rates in rural Momase revealed by the HIES are the highest in the country. Of all the major urban centers, Lae has 50.0 Figure 3: Urban rates of disputation the highest rate of disputes and Mt Hagen the lowest, followed by the 40.0 National Capital District. Lae is the Dispute rate (%) capital of PNG’s largest province 30.0 Morobe and the country’s second city, the center of manufacturing activity in 20.0 the country and its most active port. It 10.0 is an industrial hub that is in the process of rapidly urbanizing. 0.0 NCD Mt. Goroka Lae Buka Other Hagen urban Which disputes result in death? While there is no category of “homicide� in the crime incidence questions, there was a question in the disputes module which asked about whether a “dispute� resulted in property damage, injury or death. We tried to use this question as a proxy for homicide, knowing how imperfect an exercise this is. We have not reported on the rate of disputes resulting in death because these results are not robust: because of the communal nature of land ownership and inter-ethnic conflict, it is likely that a number of households within each sampling unit reported the same incident and thus amplified the numbers. Yet, 9 There are low levels of reported disputes over “distribution of aid�, “access to development resources� and “compensation associated with a development program� (all 1%), although urban residents are many more times likely than rural residents (4.83% to 0.37%) to identify disputes relating to “access to development resources� as having the most serious impact on them . 8 the data does provide some interesting insights into the type of disputes that result in deadly violence and their spatial distribution. Existing police data and Figure 4: Per cent of deaths, by type of dispute victimization surveys suggest 45 40 that the overwhelming bulk of 35 homicides occur in urban 30 25 areas.10 As outlined above, due 20 15 to the fact that rural death 10 rates are likely to be inflated by 5 0 the numbers of communal disputes, we cannot make any confident assertions about the relative prevalence between rural and urban areas. Thus, we cannot say that the HIES findings challenge the existing police data and surveys. However, the results seem to confirm the qualitative literature about localized inter-ethnic violence, which documents the “upsurge in ‘tribal’ violence across much of the Highlands over the past two decades�11 and the fact that hundreds of people are killed every year in tribal fights.12 This survey suggests that inter-tribal fighting and land disputes in the Highlands and Momase regions are responsible for a large numbers of deaths. “Other tribal disputes� would appear to be the most lethal category of dispute -- just under 3 percent of households report that their household was involved in a tribal dispute, and 50 percent of those tribal disputes lead to the death of someone. The bulk of these deaths occur in the Highlands region. Just under 1 percent of reported domestic violence cases lead to death. How do citizens resolve disputes? Over half (56 percent) of all respondents seek assistance to resolve the dispute that had the greatest impact upon them. Community leaders (32 percent), village courts (23 percent) and wantoks/family (18 percent) were the most utilized mechanisms for resolving the most serious dispute. Respondents report choosing these dispute resolution mechanisms predominantly on the basis of community respect and the power that these authorities hold. 10 In the official statistics from the Royal Papua New Guinea Constabulary (RPNGC) on homicides in 2010, Lae (66/100,000) and NCD (33/100,000) reported significantly higher rates than any other location(Lakhani and Willman 2012). Authors cross reference RPNGC data with other studies, to illustrate exactly how low the official reporting rate is likely to be. 11 Dorney (1990, 310); Reilly (2008, 15); Clifford et al. (1984, 94-5). 12 Tribal fighting in Southern Highlands province in November 2013 received considerable coverage (“Tribal fighting in PNG highlands was 'unspeakable violence'� 2013; “PNG tribal fight deaths angers local leader� 2013). 9 Wealthier respondents are more likely to use village courts, which play a pivotal role in the bulk of intra-family disputes and domestic violence. Indeed, the wealthiest quartile of respondents is the least likely to seek help from their wantoks. This is consistent with the comparative literature which suggests that wealthier citizens are generally more likely to have access to a broader range of options (whether formal courts or other institutional mechanisms or access to political or bureaucratic actors), while the poor must often rely on the most informal and localized of mechanisms. Women are more likely than men to use the most informal of mechanisms (family, church), but at the same time are more likely to seek the assistance of police. Interestingly, respondents are unlikely to seek assistance from church and NGO legal aid groups to resolve disputes (their assistance was sought in 4 percent and 1 percent of disputes respectively). These figures are similar to survey findings in other countries,13 but are likely to have their own set of explanations in the PNG context. Neither NGOs nor churches may be seen as having the authority to enforce binding outcomes. In the case of NGOs, low levels of geographic coverage, combined with low awareness of their services may explain this result. Assistance was sought from police in 9 percent of disputes. In those instances where assistance is sought, the overwhelming bulk of the cases remain unresolved at the time of the survey (90.6 percent). Due to the design of the questionnaire, in the event of an unresolved dispute, respondents were asked no further questions about the resolution of the dispute and their levels of satisfaction. This severely limits the ability to further interrogate this data. 13 This is similar to findings in other justice surveys (Akmeemana 2011), where NGOs were sought in approximately 1% of cases and religious actors in 3% of cases. 10 Figure 5: Dispute resolution mechanism availed - all dispute types 40.00 35.00 30.00 25.00 20.00 15.00 10.00 5.00 Men 0.00 Women IV. What Did the HIES Personal Security Module Show? Fear of crime and its effects The fear of crime significantly affects freedom of movement, especially for women. It stops 55 percent of women and 30 percent of men from walking at night and impacts more seriously on citizens in urban areas (58.7 percent) than rural (39.6 percent). After walking at night, fear most significantly curtails walking to work or gardens and using public transport, again with a significant rural-urban difference. Women are twice as likely to report that crime stops them from using public transport and a range of activities outside the home (walking to work or gardens, shopping, fetching water, allowing children to walk to school). Fear of crime has less impact on longer term activities such as investing in a house (6.3 percent of respondents). Interestingly, for a country where crime levels are perceived as extremely high, just over half the respondents (51 percent) report that crime does not stop them from “doing anything�. However, there is a sharp difference between men (63 percent) and women (39.5 percent) and between respondents in rural areas (54 percent) and those in urban areas (32 percent). Alcohol and drug-related crime is most feared nationally (presumably because of the unpredictable violence associated with it), with women fearing sexual assault at a much higher rate than men. Figure 6: Perceptions and fear of crime Men Women Rural Urban 11 % reporting that crime has decreased in their area in the 43.1 38.5 41 39.8 past 12 months % reporting that stealing was 42 41 43 30 the most common crime in area % reporting that alcohol and drug related crime was the most 33 33 31 46.5 common crime in area Alcohol and Sexual Alcohol and Alcohol and Type of crime one is most afraid drug related assault drug related drug related of crime (28%) (32%) crime (27%) crime (30%) More people perceive crime levels to be falling (40.6 percent) than rising (28.8 percent) in the 12 months prior to the survey, with more urban residents perceiving crime falling than rural. This finding broadly accords with the overall data from victimization surveys carried out in major urban areas since 2004.14 Figure 7: Most feared crime, percent of responses by incident type 35 30 25 20 15 10 5 0 What about the incidence of crime? There are significant shortcomings in the design of the personal security module which limits the utility of some of the data, including for international comparisons. The survey breaks down the crime incidence questions by whether the crime occurred personally to the respondent or to a child under 15 years of age in the respondent’s care. The survey then asks questions about a limited number of 14 See Lakhani and Willman (2012, 7). 12 categories of crime, many of which are unusual.15 These categories are neither internationally accepted in crime victimization surveys,16 nor do they closely correspond with offences in PNG’s criminal law (and thus the official police statistics). Because it asks respondents about crimes they have experienced personally, rather than household incidence, the questionnaire omits the important category of homicide. The overall crime rates recorded by the survey were so low as to raise considerable doubts as to plausibility. They are not reported in this note. However, while the magnitude of crimes appears far too low, the profile and the spatial distribution of crime appear plausible and are discussed below. Personal property theft was the most frequently occurring crime, which includes categories “personal property stolen in dwelling� and “personal property stolen outside home�. This was followed by assaults at home and outside. Robbery was more prevalent in urban areas than rural. As in many countries, home is a more dangerous place for women than outside. For instance, women report experiencing assault and sexual assault at three times the rate inside the home than outside. Respondents report that they know the alleged perpetrator in 37.5 percent of cases. International Crime Victimization surveys use the category of “contact crime� (which is an aggregate of robbery, sexual violence and assaults with force). In an effort to compare with international rates of contact crime, we aggregated figures for robbery, assault inside/outside home, sexual assault inside/outside home, un/provoked violence, and use of firearms in the PNG survey. While the overall magnitude of contact crime again appears to be far too low, the distribution captured by the survey appears credible. Urban centers have the highest contact crime rates in the country, followed by the Highlands. The city of Lae has the highest contact crime rate in the country, which is consistent with official data and successive crime victimization surveys which confirm Lae as a ‘hot spot’ for crime. Buka has the lowest rates of violent crime of the urban centers in PNG. What about the specific question on domestic violence? General household surveys often have serious limitations in eliciting accurate data about the incidence of domestic violence: its sensitive nature requires methodologies that are “respondent- centric� in order to bring out accurate and candid responses. The framing of the questions in the personal security module raise further complications. Female respondents were specifically asked whether they were “beaten� or “hit� by anyone in the household in the last 30 days. This contrasts with commonly accepted survey questions on domestic violence, which are usually of a much broader scope and include sexual and emotional (and not just physical) violence, and capture longer timeframes (such 15 “Car or truck stolen from your dwelling or yard�, “car of truck stolen from elsewhere�, “your personal property stolen at dwelling�, “your personal property stolen outside home�, “stealing with force or threat (Robbery), “assault at your home�, “assault outside your home�, “unprovoked violence�, “provoked violence (payback)�, “sexual assault at your home�, “ sexual assault outside your home�, “firearm used against you�. 16 For internationally accepted categories of crime, refer to the International Crime Victimization Surveys (UNICRI) at http://www.unicri.it/services/library_documentation/publications/icvs/data/. 13 as whether such violence has “ever� happened or whether violence has occurred in the last 12 months). The difference in scope and time frame makes comparisons between HIES and other surveys difficult. Nonetheless, an alarming 9 percent of women reported being beaten by someone in their household in the last 30 days. On average beatings occurred twice in the 30 day period, with the most common perpetrator being the victim’s husband (67 percent). The rates show little variation among demographic groups, except they are much lower amongst the 60+ age group (4 percent). Of those who report being hit, 73 percent did not seek help from anybody. Wealthier and more educated women are even less likely to report the incident to anyone. Among those women who did seek help, the most common person approached is a family member, wantok or friend (67 percent). Only 12.4 percent report the incident to police and 3.7 percent to a church leader. Are crimes reported to the police or anyone else? As in many developing countries, only a small number of citizens report crime to the police.17 Thirteen percent of crimes are reported to the police, with wealthier people being more likely to report crimes (an exception being wealthier female victims of physical violence in the home). This rate of reporting falls within the broad range found in the victimization surveys carried out in major urban areas.18 For those who do not report crimes to the police, 20 percent do not report crime because they believe nothing will be done. Notably, people in formal employment have considerably less confidence that anything will be done. Figure 8: Pathway for the most commonly experienced crime: personal property theft 38% did not report to anyone at all Of those 35% individuals who experienced suffered a crime in personal the last 12 months property theft Only 9.5% Only report crimes 9.5%report to to police. Of the 62% who do police . Most common reasons report, most report Most (i)lackreasons: common cited: lack & of evidence to a relative or of evidence & belief that (ii) belief police won’t do wantoks “police won’t do anything� anything 17 In the Bangladesh survey (Akmeemana 2011) 17% indicated that they approached the police for assistance in relation to an incident they experienced, although only 4% were interested in pursuing a police investigation. 18 For 2005, it is estimated that between 33 percent of all crimes in the NCD were reported to police and only 2.3 percent of all crimes in Lae were reported (Lakhani and Willman 2012, 8). 14 Further Interest? The full results of the HIES are publicly available at http://catalog.ihsn.org/index.php/catalog/3243. Summary tables of the national results have been published by Papua New Guinea’s National Statistics Office, in addition to a breakdown of the results by gender, age, rural/urban and the five geographical regions (Southern, Highlands, Momase, Islands and Metropolitan).19 The World Bank is also loading the data into a user-friendly data tool, which can be accessed at www.worldbank.org/xxxxx 15 Works cited Akmeemana, Sakuntala (2011), Crimes, Disputes and Pathways of Redress: A Household Survey on Citizens’ Perceptions and Experiences of the Justice System in Bangladesh, Washington, DC: The World Bank (unpublished manuscript). Allen, M.G. (2013) “Melanesia’s violent environments: Towards a political ecology of conflict in the western Pacific,� Geoforum, Volume 44: 152-161; Clifford, William, Louise Morauta and Barry Stuart (1984) Law and Order in Papua New Guinea (Clifford Report), Port Moresby: Institute of National Affairs and Institute of Applied Social and Economic Research, 94-5. Dale, Pamela, Kristen Himelein, Denis Nikitin and Angie Bexley (2010a) Trust, Authority, and Decision Making: Findings from the Extended Timor-Leste Survey of Living Standards, Justice for the Poor briefing note, Volume 5 Issue 1. Dale, Pamela, Kristen Himelein, Denis Nikitin and Angie Bexley (2010b) The Lay of the Land: Land Access and Dispute Resolution in Timor-Leste, Justice for the Poor briefing note, Volume 5 Issue 3. Dorney, Sean (1990) Papua New Guinea: People, Politics and History since 1975, Milsons Point, N.S.W. : Random House Australia. Filer, C. (2011) “New Land Grab in Papua New Guinea,� Pacific Studies, Volume 34, Issue 2/3: 269-294. Fulu, E., Warner, X., Miedema, S., Jewkes, R., Roselli, T. and Lang, J. (2013) Why Do Some Men Use Violence Against Women and How Can We Prevent It? Quantitative Findings from the United Nations Multi-country Study on Men and Violence in Asia and the Pacific, Bangkok: UNDP, UNFPA, UN Women and UNV. Lakhani, Sadaf and Alys M. Willman (2012) Trends in crime and violence in Papua New Guinea, Briefing Note 1, Washington, DC: The World Bank. “PNG tribal fight deaths angers local leader�, Radio Australia, 14 November 2013. Reilly, Benjamin (2008) Ethnic Conflict in Papua New Guinea, Asia Pacific Viewpoint, Volume 49, Issue 1:12-22. “Tribal fighting in PNG highlands was 'unspeakable violence,� Australia Network News, 13 November 2013. UNICRI, International Crime Victimization Survey, http://www.unicri.it/services/library_documentation/publications/icvs/data/. World Bank (2013) Papua New Guinea Economic Briefing: From the last days of the boom to lasting improvements in living standards, Port Moresby, Papua New Guinea. 16