U. Zvekic & A. Alvazzi del Frate-Criminal victimisation in the developing world. UNICRI publication No 55.

Title: Criminal Victimisation in the Developing World

DESCRIPTION: This publication presents results of the developing countries' component of the 1992-94 International Crime (Victim) Survey. In all, 13 developing countries were involved. The information is presented in two parts. Part 1 contains a comparative overview of the survey results with particular attention being paid to victimization rates, the socio-demographic correlates and the context in which the victimization occurred. Attitudes towards the police, crime reporting and costs which arise from victimization are also covered. Part 2 contains the survey reports from the 13 countries involved.

A fundamental requirement of a prevention-oriented approach to crime is the availability of extensive and reliable data, which unfortunately is not always to be found in either industrialised or, to an even greater extent, developing countries. Moreover, information that was available was based on official records produced by the criminal justice administration, and consequently its ways of dealing with crime known to the administration. Information on the "dark figure", citizens' direct contact with crime and criminal justice administration was not readily available. This situation prompted initiation of the surveys of victims of crime.

Victimisation survey is an excellent tool for collecting information on citizens' first hand experiences with crime and criminal justice.

From their initiation, victim surveys were mainly confined to the developed countries, where their diffusion was relatively rapid, becoming more focused and regular, while their presence in the developing world was very meagre. The First International Survey (1989) covered 15 developed/industrialised countries, one Eastern European and only one developing country. The broader involvement of developing countries in the International Crime (Victim) Survey initiated by the Dutch Ministry of Justice in 1989 was perceived as essential. In this context, in 1991 UNICRI was pleased to join the International Crime (Victim) Survey, with the particular task of survey co-ordination in developing countries. The second (1992) survey, out of a total of 30 countries involved, included 13 developing countries.

UNICRI's involvement in the survey in developing countries may be seen as a form of technical co-operation through research, also prioritised by the United Nations crime and criminal justice programme. This model has proved to be highly successful and we look forward to further developing it in the future.

The primary aim of the project was promotional and technical assistance-oriented, that is to say, to assist a number of developing countries to develop and implement victimisation surveys as an important research and policy tool. It was intended to introduce this tool and to highlight its research and policy potentials in developing countries with the expected result that it would transform from the 'one-shot experience' into a more regular and accepted research and policy endeavour. An important aspect of the project consisted in sensitising both the researchers and policy makers/criminal justice administrators to the significance, potentials and limits of the survey. Needless to say, the survey shared other well-known objectives of national level victimisation surveys in terms of information gathering on experiences with crime and its level, victimisation risk, propensity to report to the police, attitudes about police and punishment, crime prevention, and policy evaluation based on the results of the survey. It is also expected that the experience with the international survey will stimulate the development and implementation of national and local-level surveys.

The present volume is a daughter of a previous UNICRI publication on Understanding Crime: Experiences of Crime and Crime Control, UNICRI, No. 49, 1993, which emerged from the International Conference under the same title held in Rome in November 1992, and at which the preliminary results of the 1992 survey were presented. From that Conference onwards, UNICRI has carried out surveys in other developing countries, the last one being in Beijing, China in 1994. This volume intends to offer to the readership an overview of the main findings regarding the participating developing countries from a comparative perspective, as well as reports of the results for each participating country.

In a certain sense, we intend to provide an opportunity both for insights into each country's results as well as for a more global appreciation of problems of crime, victims and criminal justice in the developing world. This, we feel, is an important contribution towards increasing a rather meagre knowledge and understanding of crime in the developing world. Further analysis and future International Crime (Victim) Surveys will undoubtedly increase our knowledge in a dynamic perspective.

I would like to express my particular gratitude to the Ministry of Justice of The Netherlands and to Professor Jan J.J.M. van Dijk for their continuous support in this endeavour. I am also most grateful to the national co-ordinators from the participating developing countries for their most fruitful co-operation and the results achieved.

Rome, March 1995
Herman F.Woltring, Director

International Crime (Victim) Survey
Developing Countries

National co-ordinators

Sub-Saharan Africa
Joseph Safari, Senior Lecturer, Faculty of Arts and Social Science, University of Dar Es Salaam, Tanzania
Mathias Ssamula, Lecturer, Department of Sociology, Makerere University, Kampala, Uganda
C.M.B. Naude, Professor, Department of Criminology, University of South Africa, Pretoria, South Africa

North Africa
Ahmed El-Magdoub, Chief, Criminal Policy Department, The National Center for Social and Criminological Research, Cairo, Egypt
Hatem Kotrane, Professor, Director, National Institute of Labour and Social Research, Tunis, Tunisia
Abdelkhaleq Bchir, Research Assistant, National Institute of Labour and Social Research, Tunis, Tunisia

Latin America
Carlos Corvo, Lawyer, Buenos Aires, Argentina
José Arthur Rios, Professor, Faculty of Law, Federal University of Rio de Janeiro, Brazil
Luis Lachner, Senior Researcher, United Nations Latin American Institute for the Prevention of Crime and the Treatment of Offenders (ILANUD), San Jose, Costa Rica

Asia
D.R. Singh, Professor, Head, Department of Criminology and Correctional Administration, Tata Institute of Social Sciences, Bombay, India
J.E. Sahetapy, Professor, Department of Criminology, University of Surabaja, Indonesia
Celia Leones, Director, Crime Prevention and Co-ordination Branch, National Police Commission, Manila, The Philippines
Zhu Hongde, Project Director; former Deputy Director, Institute for Crime Prevention and Criminal Reform, Ministry of Justice, People's Republic of China
Wang Lixian, Deputy Director, Foreign Affairs Department, Ministry of Justice, People's Republic of China
Lu Jialun, Researcher, Institute for Crime Prevention and Criminal Reform, Ministry of Justice, People's Republic of China
Guo Jianan, Project Co-ordinator; Chief, Research Office, Institute for Crime Prevention and Criminal Reform, Ministry of Justice, People's Republic of China
Zhou Lu, Statistician; Director, Institute of Law, Tianjin Municipal Academy of Social Science

Pacific
Ken Egan, Director, Foundation for Law, Order and Justice, Boroko, Papua New Guinea

Chapter 1
Rationale and history of the project

For a long time the only available information on crime and criminal justice was that made available through official records of reported/registered cases and/or suspects and charged, tried and sentenced offenders. It is rightly claimed that these are data about the administration of criminal justice or, at the most, the citizens' action in terms of reporting incidents to the police. No doubt, such information is important in determining the volume and filtering in criminal justice administration activities. However, this information provides only for an internal measure of the functioning of the criminal justice administration. There is no measure of its effectiveness in terms of criminal activities that take place but are not in some way made known to the administration. Nor does it provide any indication of the propensity of citizens to report incidents, or some useful information on the repercussions of victimisation and attitudes towards crime and criminal justice. Official criminal justice records reflect the logic of a self-sustained and self-referential system of justice and administration thereof with offenders as prime clients. The neglect of victims is not to be blamed on official recording systems; rather it is a reflection of the administration and offender-centred criminal justice system. The growth of interest in victims of crime, in community centred crime prevention, and in accountability of police service has led, among other things, to the launching of surveys of victims of crime in a number of industrialised countries.

Differences in rates of crime over time and space are the results of differences in patterns of reporting, detecting and further processing of cases/offenders. To this should be added differences in legal definitions, concepts and measures of crime amply illustrated by problems that both the United Nations Surveys of Crime Trends and Operations of Criminal Justice System and INTERPOL data face. These problems are particularly important for international comparisons.

Experience gained in measuring crime levels and related issues on a national level through the victim surveys prompted the organisation of the international survey with a standardised methodology. This, it was hoped, would provide for independent, credible and comparative information on victimisation experiences, the context of crime, the attitudes towards crime and criminal justice policy, and an opportunity for developing and testing criminological theories in a wider context.

Though the survey has limitations, it nonetheless gives an alternative comparative perspective to statistics of offences recorded by the police, which will reflect the amount of crime victims draw to the attention of the police, and differences in police procedures as regards what offences are counted, and how. It also offers better comparative material than results from independently organised national victimisation surveys, where differences in design seriously compromise comparisons.

Some limits of the International Crime (Victim) Survey should be recognised however. It is well established that victim surveys are prone to some form of response error, to do mainly with the frailty of respondents' memories, their reticence to talk about their experiences as victims, and their failure to realise an incident that may be relevant to the survey. These factors probably mean, on balance, that the International Crime (Victim) Survey undercounts crime; it certainly means that the survey measures public perceptions of crime as expressed to interviewers, rather than 'real' experience. The critical issue here, of course, is whether response errors are constant across country. Many may be, though it cannot be ruled out that there are different thresholds for defining certain behaviours as crimes, and for wanting to talk to interviewers about these. Although survey administration was centrally organised, data collection could have differed across country, affecting what respondents were (and were not) prepared to tell interviewers. Response rates were variable, and low in some surveys. This may produce unknown effects on the results although, on the other hand, measured victimisation levels do not relate in any clear way to response rates.

A first proposal to organise an international victimisation survey was launched by the Organisation for Economic Co-operation and Development in the 1970s. Pilot studies were carried out in the USA, The Netherlands and Finland. Further to a meeting of the Standing Conference of Local and Regional Authorities of the Council of Europe held in 1987 in Barcelona, a working group was created and started developing the survey methodology and questionnaire. Some twenty countries were invited to participate in a standardised victimisation survey.

The questionnaire was conceived to cover eleven types of victimisation which could have affected the respondent personally or his/her household. A first group of crimes dealt with the vehicles owned by the respondent or his/her household: theft of car, theft from car, car vandalism, theft of bicycle, theft of motorcycle. A second group dealt with the household itself: burglaries and attempted burglaries. Finally, a third group of crimes referred to the respondent personally and dealt with robbery, theft of personal property and assault/threat. Women were also asked about sexual incidents.

In addition, the questionnaire included sections on reporting to the police and reasons for not reporting, respondents' opinion of police work, fear of crime, crime prevention measures and attitudes to punishment.

The first International Crime (Victim) Survey was carried out in 1989 with the CATI (Computer Assisted Telephone Interviewing) technique in fourteen industrialised countries. Only in Spain was part of the sample interviewed with the face-to-face technique. The sample size ranged from a minimum of 1,000 respondents in Switzerland to a maximum of 5,274 in Germany, the average size being 2,000.

At around the same time, comparable surveys were also carried out in Japan, Indonesia (Surabaja) and Poland (Warsaw). Nevertheless, the main comparative analysis focused on the first group of fourteen countries, since in Poland, Indonesia and Japan the face-to-face interviewing method was adopted and the Surabaja sample was very small (600 respondents).

Further to the 1989 survey, two needs were clearly perceived: a) to repeat the survey in the countries which had already participated in the 1989 project, and b) to carry out the survey in more countries, including developing and Eastern and Central European countries.

Within the latter context, UNICRI was contacted to explore the feasibility of victim surveys in developing countries. Following preparations in 1991, the bulk of the second sweep of the International Crime (Victim) Survey was carried out in 1992 in the industrialised, Central and Eastern European, and developing countries. The work was co-ordinated by an International Working Group.

As regards developing countries, in 1992 city surveys were carried out in Bombay, Buenos Aires, Cairo, Dar Es Salaam, Kampala, Manila and Rio de Janeiro. In Indonesia the survey covered the main cities of major islands; in Papua New Guinea, the three largest cities (Port Moresby, Goroka and Lae) were involved; in Costa Rica, the central region of the country comprising the main urban area (Metropolitan) and a mixed urban/rural area. In 1993, further to pilot studies carried out in Greater Pretoria and Tunis in 1992, surveys were also carried out in Johannesburg and Tunis. Finally, in 1994 a survey was carried out in Beijing, China.

Table 1 - list of all the sites in which the IC(V)S was carried out in the period 1989-1994

The main results of the 1992 IC(V)S were made available at the Rome International Conference on Understanding Crime: Experiences of Crime and Crime Control, organised by UNICRI in November 1992 and attended by researchers and policy makers from more than 30 countries involved in the survey. The proceedings of the Conference, including the results of the survey at country, regional and international levels, were presented in a UNICRI publication under the same title.

Further work on data analysis from the international comparative perspective is underway. The results of the IC(V)S were presented at a number of international fora. UNICRI intends to continue with the promotion of victim surveys in developing countries, and a third round of the International Crime (Victim) Survey is envisaged for 1996.

Objectives

The main objectives of the International Crime (Victim) Survey in the developing world were:

1) Promotion of crime and criminal justice information for informed decision-making.

Crime prevention and criminal justice policies cannot be developed, managed and evaluated in the absence of comprehensive, reliable and timely information. Criminal justice information is also essential for the rational allocation of scarce resources, efficient, effective and equitable administration of justice, to facilitate research, to inform the public and to forecast future developments and trends. Therefore, development and policy/research use of victimisation surveys is part of a wider process geared towards promoting conditions for informed decision-making, monitoring and evaluation of criminal justice and crime prevention in view of community needs and priorities. A notable lack of experience with victim surveys in the developing world prompted providing assistance to scholars and criminal justice policy makers/administrators to appreciate the potential and limits of the victim survey methodology and its use both for policy and research purposes.

The IC(V)S is often a part of a more complex UNICRI involvement in technical co-operation with developing countries. For example, a needs assessment mission to Papua New Guinea undertaken by UNICRI in September 1993 recommended the carrying out of local crime (victim) surveys and the creation of a Bureau of Crime and Justice Statistics, which should also have certain responsibilities regarding crime and victim surveys. The experience gained through the 1992 IC(V)S in PNG represented an extraordinary tool for sensitising the local authorities by providing them with a set of reliable data and geared towards formulation of crime prevention strategies.

Another example is the project carried out by UNICRI in China. The International Survey questionnaire was submitted to a sample of 2,000 respondents in Beijing in 1994. A Seminar on Development and Policy Use of Criminal Justice Information in China took place in Beijing from 12 to 16 September 1994 to review strategies and sources for the development of reliable and operationally valid criminal justice information. The results of the survey in Beijing were discussed in terms of their implications for crime prevention policy and improvements in the conduct of future surveys. This technical co-operation project provided opportunities for research, training and capacity building in the field of criminal justice information and its policy relevance.

The methodology adopted in most of the participating developing countries envisaged the creation of a national research team to develop the sample and train the interviewers, under the guidance of a local co-ordinator appointed by UNICRI in each country. This aspect also represents an important contribution to the enhancement of local capacity for research.

Finally, the International Crime (Victim) Survey provides a unique set of comparable information and 'constitutes a quantum leap in international statistics on crime and justice issues'.

2) Providing information on crime and criminal justice in the developing world and the development and testing of criminological theories.

The expansion of the IC(V)S to the developing countries represents an important initiative towards the provision of accurate and comparable information on crime and criminal justice in the developing world. Lack of information from the developing world has been recognised for a long time. This, in particular, relates to systematic information as highlighted from the United Nations Surveys of Crime Trends and Operations of Criminal Justice System. An analysis of the results of the four UN Crime Surveys covering the period 1970-90 highlighted that, although there was a constant increase in the number of participating developing countries in the Survey, both the quantity and the quality of information provided are still, on average, far behind that from industrialised countries. There is also insufficient information on criminological issues, research and theory development in the developing world. In addition to comparison, information based on international surveys provides ample opportunities for research development and theory building and testing for secondary analysis on, for example, the differential risk, durable and non-durable repercussions, attitudes towards police services, victim assistance and more generally crime and criminal justice policy, as well as for research development and theory-building and testing both on the local and macro levels.

Methodology

Organisation of the survey

The organisation and implementation of the survey in developing countries consisted in a number of steps which were deemed particularly important, considering the pioneering character of the endeavour.

The International Working Group first selected the participating countries, taking into account, in addition to geographical representation, available expertise at the country level. Consultations were sought with the regional institutes affiliated with the United Nations. National co-ordinators were appointed by UNICRI on the basis of their experience and involvement in previous research projects. A first meeting with the experts from seven developing countries took place in Rome in March 1991. During the meeting, the experts were briefed on the survey's objectives and methodology and requested to carry out a pilot study on a sample of 100 respondents. Further to the carrying out of the pilot study, a second meeting was held in November 1991, during which a final adjustment of the methodology was achieved, and the initial list of crimes was expanded to include consumer fraud and corruption.

While for all the developing countries, the International Working Group had contacts only with national co-ordinators either in Rome or on the occasion of international meetings, field visits were made in Indonesia and China. In Indonesia, a UNICRI/Dutch Ministry of Justice mission took place in July 1992 during which contacts were made with the national co-ordinator and the local team in charge of the survey in the various cities. In China, thanks to the support of the Government of Canada, expert advice was provided to a local research institute.

In most of the participating developing countries the national co-ordinator created an ad hoc research team usually drawn from the local university or research institutes, to develop the sample and to train the interviewers; only in Costa Rica were the services of a specialised opinion-poll company used.

Data collection

The International Crime (Victim) Survey in developing countries utilised the standard questionnaire administered by the interviewers in a face-to-face interviewing of a sample of respondents.

Some slight differences between the face-to-face (F/F) and CATI version, and between the various translations of the questionnaire into local languages, remained but it was felt that these would not impair the coherence and comparability of the instrument and the results. The questionnaire was prepared in English and translated and administered in the local languages, as indicated in Table 2.

Table 2: Overview of participation, methodology and languages used in the International Crime (Victim) Survey in developing countries

Certain difficulties with data collection were reported due to a particular sensitivity to some issues in certain cultural settings (sexual incidents; gun ownership; crime prevention devices, etc.). Problems were also encountered with respect to income, either because the respondents did not want to disclose this information or because they were unable to provide it. In some countries it was difficult to contact the respondents in the high income/residence areas because of rigid security measures surrounding their dwellings; in others, in order to facilitate access to slums (in which residents developed a system of tight control against potential penetration by the police or hostile gangs) prior contacts were established with local leaders, the neighbourhood associations, church, etc. National co-ordinators reported that in two countries refusal to respond was related to the fact that the survey was carried out on behalf of the United Nations which, in the opinion of some respondents, was biased towards their own or neighbouring country.

Sampling

One of the most serious problems faced in carrying out the survey in the developing countries was related to sampling (Table 2). The probability of sampling error is high due to the small sample size. First of all, it should be pointed out that it was decided to carry out surveys on a city level. A host of factors influenced this decision. In particular it was felt that a lack of systematic information needed for drawing a national sample was somewhat easier to overcome on a city level. It was also assumed that cities in the developing world share certain structural characteristics to a larger extent than countries to which they belong. Restricted financial means also played an important role in making this decision, particularly with respect to the transportation costs involved in a field survey.

The main parameters for sampling consisted in: residential area status, gender and age. While in some countries the sample was drawn on the basis of available census data, in others they were 'corrected' on the basis of information drawn from sociological studies and the experience of the research team. It should be noted that in some developing countries a random stratified sample was drawn through a random walk procedure.

Structure of the work

The bulk of the present volume (Part Two) contains reports on the surveys of victims of crime carried out in thirteen developing sites. They are meant to provide rather detailed information and meet the interests of readers who are particularly keen to learn about a specific locus under study. They are, in a certain sense, repetitive in format and structure but, indeed, they are not expected to be read in sequence. An attempt was made to provide some additional information in order to assist the reader to better appreciate the context in which the results of the survey are placed. Yet this is not sufficient for a full cultural appreciation of the context or the results of the survey. Recognition of the cultural relevance and the influence of socio-cultural specificities is imperative for a meaningful interpretation. This recognition indicates both the potentials and limits of the International Crime (Victim) Survey.

The first part of the volume presents an overview of the results from a comparative perspective regarding the developing world. It is herewith that both the editors and readers become fully aware of the potentials and limits of the survey. While pursuing comparability through a standard methodology a certain price is paid to the cultural relevance and specificity. To match the two is an issue at the heart of international comparison, and there is no easy solution. Although standardised methodology provides an enormous advantage for comparisons, it also provides an enormous disadvantage for culturally-meaningful appreciation. A desire to achieve internationally comparable and standardised information on crime and justice raises some crucial conceptual and methodological issues. This, in particular, regards the assumption of universality which applies to the public's non-legal concepts of crime as well as to the official legal definitions. The International Crime (Victim) Survey has put this assumption to an empirical test by designing, piloting and administrating a questionnaire covering the common element in the public's experience of crime in a wide range of geographical, cultural and legal contexts. The experience gained with the two sweeps of the IC(V)S was encouraging. The concepts of certain conventional crimes may indeed be shared by the public; certainly more than the legal definitions. Lastly, it should be appreciated that the main objectives of the IC(V)S were: the collection and analysis of comparable information at the international level, and the promotion of the survey as an important policy and research tool in countries lacking experience in this area, such as developing countries and countries in transition. It is hoped that, based on this experience, further development and use of more culturally-tailored surveys on a national/local level will follow. This is particularly important in countries which lack developed criminal justice information mechanisms and structures.

Chapter 2
An Overview of Main Findings

Context

The International Crime (Victim) Survey provides a wealth of empirical data in a cross-cultural perspective. In this perspective it is possible to assess differences in national and local crime and victim profiles, and relate them to social, economic and cultural environments. It can assist in the search for regional patterns and confront them with those identified on the basis of administratively-produced measures of crime and operations of the criminal justice system.

One of the limits of the analysis presented herewith consists in the fact that data that is currently available represent the fixed point in time in which they were collected for the first time. For a more complete analysis, it would be important to have the possibility to examine trends in both development and crime.

The forthcoming 1996 IC(V)S will provide the opportunity for such an approach, and many issues which are difficult to understand might find a better explanation once a set of longitudinal data becomes available.

In this chapter, any reference to comparison with industrialised and/or Central and Eastern European countries is only illustrative. This was intentional so as to keep the focus exclusively on developing countries and to highlight issues and problems pertinent to the developing world. Chapter 3 provides selected summary notes in the global comparative perspective, while an elaborated analysis at the comparative level is underway and will soon be made available.

Developing countries participating in the survey, in terms of their population size, range from Costa Rica and Papua New Guinea (3,190,000 and 4,056,000 respectively) to China and India which, with populations of more than 1 billion and 880 million respectively, rank as the first and second largest countries in the world. Cities participating in the survey also differ along a number of indicators such as size, rate of growth, etc. They also differ substantially on a number of developmental indicators of economic performance, urbanisation, human development index, etc.

Table 1: Data on developing countries/cities covered by the International Crime (Victim) Survey

For example, in terms of Human Development Index, countries range from Argentina (high), to Uganda, Tanzania, Papua New Guinea and India (low).

Other issues should be taken into account, such as income disparities. In Brazil, for example, the ratio between the income share of the bottom 20% of the population and that of the top 20% is 1 to 32. In South Africa the disparity between blacks and whites is four times larger than in the United States and if separate Human Development Indexes were calculated, that for whites would be 0.878 and would rank between Spain and Hong Kong (high), while that for blacks would be 0.462 and would fall at the opposite end of the human development scale.

Rapid population growth is included among those indicators which are generally considered to be crime-generating. 'The real threats to human security in the next century will arise more from the actions of millions of people than from aggression from a few nations - threats that will take many forms:

• Unchecked population growth
• Disparities in economic opportunities
• Excessive international migration
• Environmental degradation
• Drug production and trafficking
• International terrorism

In this respect, Table 2 presents some of the indicators that might provide a framework for interpretation.

Table 2: Population growth rate, crude birth rate, total fertility rate and urban population growth rate in thirteen developing countries

Source: World Bank, 1993 and UNDP, Human Development Report 1994 - UNICRI elaboration

It can be observed that a rapid growth of both the global and urban population, accompanied by the highest crude birth rates and total fertility rates, are found in two countries from Sub-Saharan Africa: Tanzania and Uganda. Dar Es Salaam and Kampala, the capital cities of these countries, were actually the most violent among those covered by the survey. 'In terms of aetiology of crime, the phenomenon of massive rural/urban migration becomes significant when the migrant group consists mainly of young able bodied citizens, usually with only the first level (primary) education, and mostly in the juvenile age group. Emigration atomises the family thereby neutralising the youth's most fundamental traditional social control bonds. (...) When a youth is viewed in the context of the ecological theories of crime causation and the delinquent subculture theories, operating within the framework of the theories of the broken homes, differential association, economic conditions, culture conflict and anomie, it is clear that he is a potential recruit for the rank and file of the criminal population'. However, it can be noted that in China the very high urban population growth rate is not accompanied by a high crime rate.

Needless to say, there are a host of social processes and factors influencing, to varying degrees and in diverse directions, distinct regional and or national/city crime patterns and rates. No attempt was made to analyse them and the presentation that follows in the next section is purely descriptive.

Victimisation rates

This section presents an overview of the selected main findings in developing countries. The victimisation rates presented are prevalence rates: the percentage of respondents aged 16 or more who reported being victims of crime once or more, either individually or as members of a household. The time span covers the calendar year preceding the survey (in most cases 1991; in South Africa and Tunisia 1992 and in China 1993).

In order to facilitate the analysis, countries/cities have been grouped according to the geographic region they belong to. As a result, five regions have been identified as follows: Sub-Saharan Africa (Kampala, Dar Es Salaam and Johannesburg); North Africa (Cairo and Tunis); Latin America (Buenos Aires, Rio de Janeiro and Costa Rica); Asia (Beijing, Bombay, Jakarta and Manila) and Asia/Pacific (comprising the Asian cities plus Papua New Guinea which includes cities of Port Moresby, Goroka and Lae).

An overall view into one-year victimisation experiences is presented in Table 3.

Groupings of individual crimes and their regional distribution are presented in Figure 1. Vehicle crime includes theft of car, motorcycle and bicycle; property crime refers to theft from car, burglary and attempted burglary, theft of personal property and car vandalism; violent crime: robbery and assault/threat.

[ Figure 1 ]

Property crime is the most frequent form of victimisation among those considered in Figure 1. It largely affects Sub-Saharan Africa (more than 50%) and, to a lesser extent, the other regions. The second most frequent form of victimisation in all regions is violent crime, with the exception of Asia. It should be noted that Asia shows the lowest rates for all types of crime. In comparative terms, the highest victimisation rates for all those types of crime were observed in Sub-Saharan Africa.

Consumer fraud/corruption

Data from Table 3 reveal that, among the types of crime covered by the survey, consumer fraud and corruption stand out as the most frequent in all the regions of the developing world and in most of the cities within the specific region.

[ Figure 2 ]

Differences in levels of exposure to consumer fraud vary somewhat more across the regions with respect to corruption. Both Sub-Saharan and North Africa lead in consumer disrespect, as they do together with Latin America in regard to how things are achieved when dealing with public officials. Asia, with and without Papua New Guinea, appears to be a less corruptive public environment. In Papua New Guinea, which otherwise inflates the overall victimisation level of Asia, neither consumer fraud nor corruption are the most frequent victimisation experiences.

The way this particular victimisation experience was measured provides information on what one might call 'elementary level' victimisation. In other words, it regards those types of fraud and corruption to which an ordinary citizen is, as we know by now, very often exposed to by the commercial and public sector. Consumer fraud regards cheating in the quantity and quality of goods attained and services received. Corruption refers to bribes requested by public officials. Consumer fraud and corruption speak about development itself, about the citizen's position vis-à-vis government and the commercial/service sector, the lack of consumer/client and citizen protection, and the ways in which people go about, or are made to go about, in satisfying their needs and rights. There is no doubt that the developing world experiences also advanced as compared to the 'elementary' forms of both fraud and corruption. However, the relationship between the elementary and advanced forms is perhaps analogous to that between the organised and conventional crime: the more they go hand in hand, the less easy it is to draw a definite borderline between the two. Nor should one conceive of elementary forms as a prelude to more advanced ones. Yet, for an ordinary citizen the likelihood of a first hand victimisation experience with the elementary forms is higher than with the advanced ones both for consumer fraud and corruption.

Property crime

Urban areas of the developing world experience rather high rates of property crime as shown in Figure 3.

[ Figure 3 ]

At the regional level, there is again Sub-Saharan Africa with the highest rates for all crimes on the one side, and Asia with the lowest rates on the other. In terms of individual property crimes, the most frequent is exposure to theft of personal property. Theft from car and attempted/completed burglary are least frequent in Asia, although with the inclusion of Papua New Guinea the rates come close to those of Latin America. At the global comparative level, Africa in particular and Latin America to a lesser extent have higher rates than the developed world.

High risks of theft of personal property are marked for citizens of Kampala, Dar Es Salaam and Port Moresby, Goroka and Lae (Papua New Guinea), and this is also the case for burglary.

The most frequent form of vehicle theft is theft of bicycle, followed by theft of motorcycle and theft of car.

[ Figure 4 ]

Victimisation rates for vehicle theft can be expressed in two forms: as rates of the total sample and as rates for vehicle owners. The former are most useful for comparing the spread of different victimisation experiences in the population; the latter provide more information about the real risk of the target group, that is, the vehicle owners.

Theft of car is most widespread in Buenos Aires which also has the highest car ownership rate, while, at the regional level, the highest rate is in Sub-Saharan Africa with the second highest ownership rate. This finding supports the opportunity theory according to which the higher the ownership rate the higher the victimisation rate.

Theft of bicycle is most frequent in Beijing and the urban areas of Papua New Guinea which makes Asia (with and without Papua New Guinea) the region with the highest propensity for bicycle theft, although similar levels are recorded in Latin America and Sub-Saharan Africa.

The vehicle owner's risks are of course higher. This in itself is a trivial finding. Figure 5 reveals that car owners' risks are highest in Sub-Saharan Africa and are then quite similar in Latin America, North Africa and Asia/Pacific.

[ Figure 5 ]

On average, the owner's risk is at least twice as high as the population spread of car theft with the exception of Latin America. These differences indicate certain limits of the target-availability explanation and introduce the unequal ownership structure and scarcity as a necessary part and parcel of the explanatory paradigm. Indeed, within the demand-supply paradigm as developed by van Dijk, victims are seen as reluctant suppliers. This supply role is particularly important in the situations of unequal ownership structure and scarcity and goes hand in hand with demand.

An important indicator of the nature of theft of car and the efficiency of law enforcement is provided by data on car recovery rates. These rates are presented in Table 4. On a country level, the highest recovery rates are in the cities of Papua New Guinea, Costa Rica and Africa. Strong doubts exist as to whether these data indicate the effectiveness of the law enforcement although, at least in the case of PNG and Costa Rica, the size of the territory and, in particular for PNG, restricted opportunity for traffic (due to a restricted road network) can ease the recovery of the stolen vehicles. There are valid arguments that a high recovery rate indicates that vehicles are stolen for joy-riding purposes, an explanation offered for industrialised countries. The attractiveness of cars for young people, particularly in a situation of relative car scarcity, supports the joy-riding hypothesis. The recovery rate in the developing world is much lower than that in the industrialised world which, comparatively speaking, indicates both lower levels of efficiency in recovery and less joy-riding. In addition, an examination of data from Europe suggests that low recovery rates may indicate the existence of an organised crime network smuggling cars from Western Europe to Central and Eastern Europe, thus not only reducing the recovery rate but also increasing car theft rates both in Western and Central and Eastern European countries. The plausibility of extending this hypothesis to the developing world comes from INTERPOL data, according to which more than 1.6 million cars were stolen across Europe in 1991, of which 633,000 were never recovered and, furthermore, '...more than half of these cars were stolen in Britain, so it is logical that they ended up in countries that drive on the left-hand side of the road'. Since the majority of left-hand side driving countries are developing countries (in addition to Australia and Japan), these data indicate the involvement of organised crime and its influence on increasing car theft rates in both the developed and developing world, as well as a reduction in recovery rates.

Violent crimes

Violence as represented by robbery and assault/threat is much higher in Sub-Saharan Africa compared with other regions of the developing world. Latin America also exhibits a high level of violence.

[ Figure 6 ]

The often indicated low level of violence in Asia is confirmed by these data, as well as for North Africa in the case of assault/threats. Cities in Papua New Guinea stand out with high levels of violence (assault/threats), followed by the African cities of Kampala, Dar Es Salaam, Johannesburg and Latin American: Rio de Janeiro and Buenos Aires. Tunis and all the Asian cities are at the lower end of the violence scale. As regards robbery, Rio de Janeiro, the PNG cities and Dar Es Salaam exhibit particularly high rates. From a global comparative perspective, with the exception of Asia, the developing world cities exhibit at least three times higher robbery rates as well as higher assault/threat rates than urban areas of the industrialised world.

This summary overview of victimisation rates highlights that, in most developing countries, cities are at a high risk of victimization. Regional comparison evidentiates that the highest level of victimisation is in the Sub-Saharan Africa, while the lowest is in Asia. A host of social, economic, demographic and cultural factors contribute to regional victimisation profile.

In particular, the survey reveals high level of property crime, including vehicle-related crime. It is interesting to observe that one of the objections raised by local experts at the time of discussing the survey questionnaire referred to it as being too "western" and too much focused on vehicles and other property. However, the survey results show all the relevance of vehicle and frequently related victimisation in the developing world, too.

Environment and costs of crime

Area characteristics and fear of crime

Respondents were requested to provide an assessment of the social cohesion of the area in which they resided. The regional distribution of the responses is presented in Figure 7. Two patterns appear very clear: on the one hand, Sub-Saharan Africa and Asia (with and without PNG) show a predominance of respondents who describe their area as one where people mostly help each other; on the other hand, North Africa and Latin America have higher percentages of respondents who declare that in their area people mostly go their own way, or that there is a mixture of the two.

[ Figure 7 ]

The fact that the majority of respondents describe their area as very cohesive does not mean that they feel more secure when walking in the same area after dark. In fact, as is shown in Figure 8, the two regions where the highest percentage of respondents admitted feeling 'very safe' after dark are North Africa and Latin America which were described by the respondents as the least cohesive.

[ Figure 8 ]

Nevertheless, if both the 'very safe' and 'fairly safe' categories are considered, the safest regions are Asia and North Africa, while the highest levels of insecurity were registered in Sub-Saharan Africa (with a peak in Johannesburg where these data might not only indicate fear of crime but also other types of insecurity) and Latin America (especially Rio de Janeiro).

As can be observed in Figure 9, there is some correlation between the rates of assaults/threats and robberies and fear of crime. The Figure shows that the higher the crime rates are, the higher is the feeling of insecurity (Sub-Saharan Africa and Latin America). However, high rates of precautions adopted when going out after dark (avoiding certain areas that are considered dangerous and asking somebody for company) do not necessarily mean that respondents feel unsafe, and are also adopted in the low-crime regions. They may also reflect the cultural pattern of an out-going style which is not directly related to fear of crime.

[ Figure 9 ]

Place of victimisation

Survey results show that personal victimisation occurred more frequently far from home, elsewhere in the city. Figure 10 shows distribution of places where the various incidents happened. For robberies, personal thefts, sexual incidents and assaults/threats the highest percentages registered refer to another area of the city. The lowest rate for incidents which happened near the respondent's home was observed for personal theft.

[ Figure 10 ]

The majority of vehicle-related crimes occurred in the area of residence. It appears that these crimes are mostly nightly endeavours, especially thefts of bicycles. Approximately one-third of the incidents related to vehicles happened elsewhere in the city. It should be noted, however, that facts related to vehicles fall into the category of household victimisation and were described with less precision by the respondent, who might ignore some details of the event.

However, criminal victimisation does not only occur out in the street. Among the types of crime covered by the survey, violent crimes often happen in the victim's own house or in its vicinities. Domestic violence is a frequent event and the survey results can provide an indicator of its seriousness. For sexual incidents and assault/threat the category 'in own home' was provided: the response distribution by country is presented in Figure 11. Approximately one-third of the cases of assault in Kampala and Manila and of sexual incident in Kampala took place in the victim's own home. In most countries, at least one out of six aggressions occurred in the respondent's home.

[ Figure 11 ]

Further information in the field of domestic victimisation can be obtained through the analysis of the offender's description in cases of assaults/threats (Figure 12). More than half of the assaults/threats were committed by offenders known by the victim, at least by sight.

[ Figure 12 ]

People from the family or very close to it (partners/ex-partners, relatives or close friends) are involved in 20% of the assaults. In such cases the survey provides an insight in household life with the disclosure of events that would never be reported to the police. It is also possible that, if reported, some of those events would not be considered "crimes" and, therefore, the survey might overcount "crimes". This in particular relates to domestic violence, sexual incidents included.

Weapons involved in committing crimes and physical injury

Assaults/threats were, in the majority of cases, committed without the use of a weapon (Figure 13). A greater use of weapons can be noted in most cases of robberies in all regions with the exception of North Africa. In this respect, Sub-Saharan Africa and Latin America appear as the most violent regions, in particular due to the frequent use of weapons in both robberies and assaults/threats in the cities of Dar Es Salaam and Johannesburg and especially in robberies in Buenos Aires and Rio de Janeiro.

[ Figure 13 ]

As a matter of fact, regions with higher violent crime rates and higher use of weapons in committing crimes also show a higher percentage of physical injury suffered as a consequence of the assault. Physical damage was reported by approximately a third of the victims of assault in Sub-Saharan Africa, and a quarter in North Africa and Latin America. Lower percentages declared having recurred to medical treatment in all regions. In particular, the lowest percentage of victims who sought medical care after an aggression was observed in Latin America, despite high level of violence and usage of weapons. This data is consistent with low reporting rate for assaults observed in Latin America.

[ Figure 14 ]

Household protection

A series of crime prevention measures are adopted by the respondents to protect their households. The most popular are door locks and grills on doors and windows which are available to almost half the respondents in all regions. Although some regional differences can be observed (Figure 15), smaller percentages of respondents keep watchdogs and have high fences (especially in Sub-Saharan Africa and Latin America). A few respondents affirmed that their house has a caretaker or a burglar alarm. Finally, neighbours watch in case of absence in approximately one-fifth of cases in all regions (a bit less in Asia, however, where all crime prevention measures are slightly less widespread than in the other regions).

[ Figure 15 ]

A comparison between rates of completed and attempted burglary, a constructed index representing an average of crime prevention measures adopted and the likelihood of burglary in the next twelve months as expressed by the respondents is presented in Figure 16.

[ Figure 16 ]

In general, spread of crime prevention measures is higher than the actual crime, with the exception of Sub-Saharan Africa where more respondents were victims of burglaries than those equipped with household protection devices. On the other hand, fear of crime (likelihood of burglary) is positively correlated with victimisation. The region with the highest spread of crime prevention devices was Latin America, which also shows the highest percentage of guns owned for crime prevention purposes (15% of the respondents).

Seriousness of crime

The respondents' assessment of the seriousness of victimisation experiences shows variations depending on the type of crime. A global overview is provided in Figure 17, in which a ranking of the eleven types of crime covered by the survey is shown according to the percentage of respondents who considered that particular crime 'very serious'.

[ Figure 17 ]

As can be seen in Figure 17, on average half of the respondents consider the events they have been victims of as very serious. The most serious type of victimisation is theft of car, which also ranks first in all the regional distributions. Although car theft does not represent an immediate physical danger for the victim, it means a serious financial loss and is considered a very serious event in all cultures. Burglary, sexual incidents, robbery and assault/threat follow.

It is interesting to note that car theft and burglary are also those crimes which are more frequently reported to the police. In addition, many respondents have their cars and/or households insured. The respondents are therefore quite concerned about the goods in need of protection and their recovery or compensation for loss.

The second group of crimes perceived as 'very serious' is that involving contact between the offender and the victim: sexual incidents, assault/threat and robbery.

[ Figure 18 ]

Figure 18 shows the two opposite patterns of seriousness assessment observed in Asia and Sub-Saharan Africa respectively. The highest levels of concern with crime seriousness is observed in Sub-Saharan Africa. There is very little differentiation between the various types of crime. This region deviaties from the average ranking of seriousness, in particular for assault/threat which is there in the bottom part of the seriousness ladder.

A much lower proportion of victims in Asia expressed such an assessment of seriousness for all types of crime. A different 'philosophy' towards victimisation is revealed by citizens of Asian cities, who make a wider use of the 'fairly serious', rather than the 'very serious', category. Car theft is the only type of crime which is considered very serious by 60% of the respondents and fairly serious by the remaining 40%. All the other types of crime are considered less serious, with theft of bicycle ('not serious' for 65% of the respondents) at the bottom of the seriousness scale.

Sub-Saharan Africa is a crime-driven region. It appears that the respondents are under pressure (see also the high indicators of fear of crime) and that frequent, repeated experiences with crime make it inevitable that this problem be considered very serious, irrespective of the type of crime. On the other hand, the lowest victimisation rates were observed in Asia, with a more differentiated attitude to the seriousness issue. It appears that the differential assessment of crime effects is facilitated in communities in which less crime occurs.

Women

Since household crimes (theft of vehicles, burglary and attempted burglary) cannot be used as direct indicators of gender-related risks, only three types of personal victimisation (robbery, assault/threat and theft of personal property) have been further analysed in this respect. Survey results indicate that in all regions women run lower risks than men for violent crimes (robbery and assault/threat).

[ Figure 19 ]

It should be noted that the risk of robbery is low in Asia and that in this region there is almost no difference between the two genders. Gender differences are generally not marked for both types of violence, with the exception of assault/threat in North Africa, where the exposure of males to this type of victimisation was almost twice as high.

On the other hand, women were more often victims of theft of personal property (they carry bags more often; they are perceived as less reactive, etc.). Higher risks for women were observed in Asia, Latin America and Sub-Saharan Africa while in North Africa, although percentage-wise women ran approximately the same risk as in the other regions, males were more exposed to victimisation by theft of personal property.

[ Figure 20 ]

Women fear going out after dark more than male respondents and often ask somebody to accompany them. In addition, a behaviour which has proven to be a women's favourite in all regions is to avoid certain areas which are perceived as being dangerous.

Sexual incidents

Problems of sensitivity to the issue of sexual incidents might have created distortions in reporting to the survey, either in the direction of over-reporting or under-reporting. In some of the most industrialised countries covered by the IC(V)S it was noted that high sensitivity to the issue corresponded to high rates of victimisation reported to the survey.

It appears that this is not the case for developing countries, where survey results have shown that lower rates of sexual harassment were found in countries where women enjoy a better status.

Table 5 shows several indicators related to the advancement of women. A negative correlation was found between victimisation for sexual incidents and adult literacy rate, years of schooling, average age at the first marriage and contraceptive prevalence rates. Vice versa, a positive correlation was found with crude birth rate.

Table 5: Selected development indicators and sexual incident rate

Data source: UNDP, Human Development Report, 1994 and IC(V)S - UNICRI elaboration.

Experiences with sexual incidents were analysed by addressing specific questions to women only. Some of the problems which are normally encountered in administering questions on sexual incidents (such as definition of sexual incidents and readiness to report them to the interviewer) were found in most of the developing countries covered by the survey. It should be stressed that the measure of sexual offences provided by the survey must be taken with care, especially as regards comparison. In fact, the wording of the question itself might elicit different answers in different languages and contexts which, in turn, might have different levels of sensitivity to the issue.

Survey results reveal that this type of victimisation is very frequent in many developing countries. Data grouped by regions are shown in Figure 21, in which both five-year and one-year victimisation rates for sexual incidents are presented.

[ Figure 21 ]

As is the case with most types of crime covered by the survey, one-year data represent more than one-fifth of the five-year corresponding rates and tend to approximate one-fourth of the five-year data. This might mean either that the respondents have a better recollection of events which happened more recently or that there is a trend towards an increase in this type of crime. However, regional comparison on the one-year rates shows smaller differences between the regions than the five-year rates.

The highest rates were found in North Africa, where 33% of the interviewed women experienced sexual incidents in the last five years, followed by Latin America, Sub-Saharan Africa and Asia/Pacific with rates close to 15%. The least victimised were women in Asia (10%).

As far as the seriousness of the incidents is concerned, three indicators are provided by the survey:

• a) a description of the incident (rape, attempted rape, indecent assault, offensive behaviour);
• b) whether the victim considers it a crime or not;
• c) an assessment of the incident's seriousness (very serious, fairly serious, not very serious).

[ Figure 22 ]

As regards the description of the incident, two levels of seriousness were considered in this analysis: incidents described as rapes, attempted rapes and indecent assaults belong to the more serious category, while others are considered sexual harassment of a less serious nature. The regional distribution (Figure 22) reveals that in North Africa, where the highest frequency of incidents was observed, more than three-quarters of the incidents were simply described as offensive behaviour. Nevertheless, in North Africa more than 60% of the victims described the incident as 'very serious' (Figure 23) and more than three-quarters declared they considered it a crime (Figure 24). This discrepancy between the three indicators of seriousness provided by the survey suggests that sensitivity to this type of incident might vary depending on local culture, the willingness of the respondents to talk about this experience and, perhaps, the meaning attached to the questions submitted.

Among the incidents reported to the survey, some were described as completed rapes. This occurred more frequently in Sub-Saharan Africa and Papua New Guinea. In addition, in Sub-Saharan Africa almost half of the victims experienced an attempted rape. This type of victimisation was also very frequent in Latin America. Many women also declared having experienced indecent aggressions in Asia and Latin America. In general, the region with the highest level of sexual offences of a serious nature was Sub-Saharan Africa, followed by Latin America.

[ Figure 23 ]

Consequently, similar proportions of victims in Sub-Saharan Africa and Latin America affirmed that the incidents they experienced were very serious, and most of them considered them a crime. In Asia, a lower percentage of women considered the incident very serious, but approximately half of them claimed it was a crime, while some 20% could/would not qualify the incident, indicating a great margin of uncertainty about the assessment of the seriousness of such events.

[ Figure 24 ]

It is particularly interesting to observe that, in all the countries, only a small proportion of the sexual incidents which were described as crimes were reported to the police, as is shown in Figure 25, indicating women's negative appreciation of and/or experience with the police. To this one should add often special relation with the offender restraining reporting to the police.

[ Figure 25 ]

Indeed, while most of the sexual incidents occurred elsewhere in the city or the country or abroad, some 40% took place either near to, or in the victims' own home (Figure 26). Women are often victims of such incidents and it was observed that most sexual incidents were not reported to the police. Although, according to survey data and other studies, women tend to report crimes more frequently, the presence of a victim-offender relationship might make many victims restrain from reporting to the police. In fact, in many cases the victim knew the offender, at least by sight. In smaller but significant percentages the offender was the partner/ex-partner, a close friend or a relative. Finally, 23% of the victims refused to provide details about the offender(s), thus suggesting they actually knew him or them.

[ Figure 26 ]

Finally, fear of violent crime is mostly expressed by women, despite the fact they appear to be at a lower risk of assault. This result is consistent with similar findings in the industrialised countries. Furthermore, women victims of crime, in particular of sexual incidents, feel less safe, are more negative about police and more punitive towards offenders.

Police, citizen's experience and attitudes towards punishment

Reporting to the police

To a large extent, it is expected that the higher is the value attached to the target of criminal activity (whether tangible or intangible), the higher is the propensity to report to the police. However, this expected propensity to report is influenced by a number of factors, including those pertaining to the domain of past personal and/or otherwise acquired experience with the police, through those pertaining to the sphere of expectations, to those more closely related to some special properties of the victimisation experience. Some are related to the relationship between the police and citizens (personal or stereotypical esteem and confidence); others to the existence of alternative ways and means to deal with crime and victimisation; still others with the nature and perceived seriousness of the victimisation, or the relationship with the offender, or the 'privacy' of the issue at hand.

Table 6: Reported crimes (percentages)

These data present reporting rates for selected types of victimisation (crimes). As expected, the highest number of reported cases is related to car theft and the lowest to personal theft and sexual incidents. In nine participating sites more than 85% of car theft victims reported the incident to the police the last time their car was stolen; this was the case with about 70% of victims in Cairo, Costa Rica and Papua New Guinea. Being extremely scarce in Beijing, cars are of particular value and, consequently, all cars stolen are reported to the police. Bombay, with 54% of reported car thefts, confirmed its position of a city with the lowest average reporting rate for any type of victimisation. Reporting rates for theft of motorcycles are somewhat lower; it is however interesting to note that owners of motorcycles in Manila reported all cases of stolen motorcycles, a somewhat surprising finding taking into account a low-reporting practice for other vehicle and vehicle-related thefts. Theft of bicycle is in particular not reported in Johannesburg, Rio de Janeiro and Jakarta. The reporting rate for car vandalism ranges from none reported in Beijing to 73% reported in Dar Es Salaam. On average, Latin American citizens tend to report just a few cases of car vandalism contrary to the citizens in Papua New Guinea and Dar Es Salaam.

Around 70% of burglary cases were reported to the police in Dar Es Salaam and Buenos Aires, while the lowest reporting rate was found in Cairo, Bombay and Manila. Citizens of Jakarta, Papua New Guinea, Tunis and Dar Es Salaam reported around one-third of personal theft cases while less then 10% were reported by citizens of Kampala and Bombay.

Two forms of criminal violence - robbery and assault/threat - are on average reported in 30 to 40% of the cases, although as much as 66% of robbery and assault/threat cases were reported in Dar Es Salaam, and as little as 11% of assault/threat in Rio de Janeiro. It should be noted that, on average, burglary is more often reported than robbery, with the exception of Cairo and Manila.

Both types of reporting of sexual incidents (in the victim surveys and to the police) seem to reflect certain particularities of a cultural nature related to the women's position, awareness, freedom, the concept of privacy, and the gender-biassed police culture. From a comparison of the two reporting practices, two discernible patterns are found. On the one hand, there are sites in which both victimisation and reporting are high (Dar Es Salaam and Papua New Guinea) and, on the other, those in which the relationship between the two reporting rates is inverse; for example, the rate in Cairo is 10% while only 2.5% of cases were reported to the police, or, in Manila the victimisation rate is low (1.2%) while 16.7% of the cases were reported to the police.

[ Figure 27 ]

Citizen's experience and satisfaction with the police

Within the ambit of multiple factors influencing the reporting of victimisation to law enforcement bodies, particular attention is given to reasons for not reporting. Among these reasons, in turn, attention is given to the citizens' satisfaction with the way the police handles reports and consequently acts (or does not act). While reasons for not reporting differ from crime to crime it is still possible to decode the existence of certain clusters. The reasons for not reporting can be broadly differentiated in those related to the 'weight' of the event itself (seriousness of crime, lack of evidence and inappropriateness of police activity), access to solutions away from law enforcement (solved it myself, family support), and negative attitudes (experience/belief) towards police (could do nothing, would do nothing, fear/dislike). In general, the most common reason for not reporting regards the event itself, and most often its relatively low level of seriousness. This category is followed by an experienced or assumed attitude expressing the lack of effectiveness on the part of law enforcement. Self-help and/or that of the family, friends or other institutions ranks third in this general picture of factors influencing non-reporting. It is however extremely difficult to find any pattern in terms of the importance of the reasons for not reporting. There are too many variations across crime types and countries.

However, since this section deals with the police, one can look a bit more closely at the police-related reasons. 'Police could do nothing' indicates a belief that the police are helpless in their efforts to recover property, find the offender, or do something else which would be beneficial for the victim. It does not say why the 'police could do nothing'; in other words, it is an attitude of resignation on the part of citizens. On the other hand, 'police won't do anything' is a clear statement that implicitly contains a criticism of the police's deliberate reluctance to do something that could be beneficial for the victim and which should be a normal part of the police's role and expectations related to it. 'Fear/dislike of police' indicates a negative attitude towards the police, either because an involvement of the police might have unpleasant consequences for the victim (fear of reprisal) or because there is a general disapproval of the police. Particular attention should be paid to the statement 'police won't do anything' and 'fear/dislike of police'. This information should also be related to another indicator of police-community relations, namely the reasons for dissatisfaction with the way police handle cases once reported.

In all sites, 'police won't do anything' is almost exclusively related to property crimes - personal theft, bicycle or motorcycle theft, theft from car and burglary. Robbery is indicated as crime about which the 'police won't do anything' in Johannesburg, Papua New Guinea, Cairo and Rio de Janeiro. 'Fear/dislike of police' is, as expected, most frequently indicated in connection with sexual incidents. That women victims of sexual incidents are often treated badly by the police is a well recognised fact. Other crimes for which victims are reluctant to report to the police because of fear or dislike are violent crimes such as robbery and assault/threat. These three types of crime are all contact crimes and might involve particular relationships with the offenders or sometimes even a life style which may, in the police perception, result in treating a victim as 'an accomplice' or 'one who deserves what he/she got'. In addition, some property crimes are also indicated, the investigation of which might lead to unpleasant situations.

Table 8: Satisfaction with the police on reporting crimes

In the developing world there are more victims who are dissatisfied with the ways in which the police handle reported cases of victimisation. However, an extremely high no response rate is registered which complicates any analysis. It should be noted that the difference between 'favourable' and 'unfavourable' is the largest in Jakarta while it is lowest between the 'unfavourable' and 'favourable' in Kampala and Beijing.

Table 9: Reasons for dissatisfaction with the police upon reporting (*)

(*) multiple answers possible

Dissatisfaction with the way the police handled the matter once reported is another important indicator of the police-victims relationship. However, the available information presented in Table 9 is based only on the last incident reported to the police and pertains to those who were dissatisfied with the way the police handled their reports. It should be underlined once again that reasons for dissatisfaction vary by crime types. Nevertheless, some general patterns are discernible. At the regional level, somewhere around one-third of the dissatisfied victims stated that the police 'didn't do enough', indicating lack of full commitment on the part of the police. Other two important reasons stand out: the police 'didn't recover my property' and 'didn't find the offender'. While these are interrelated, the former indicates that reporting is often motivated by the expectation to reclaim stolen property. In the developing world, substantial economic interests might be at stake for the victims of property crime. This reason was indicated by more than 40% of dissatisfied victims in Johannesburg, Dar Es Salaam, Rio de Janeiro and Beijing. Inability to find the offender was indicated by more than one-third of dissatisfied victims in Dar Es Salaam, Cairo and Rio de Janeiro. As already noted, in most of these cities victims also revealed that reasons for not reporting have, to a large extent, to do with 'police won't do anything'.

[ Figure 28 ]

Nor are the citizens in the developing world satisfied with the way the police control crime in their area of residence. There are more 'dissatisfied' than 'satisfied' respondents in all the regions with the exception of Asia. There is particular dissatisfaction among the citizens of Latin America (Buenos Aires: 65% to 28% and Rio de Janeiro: 77% to 12%). In Asia, on the other hand, 83% of the citizens of Beijing state that the police do a good job in controlling crime in the area in which they reside, which is the case only with 21% of the citizens from Papua New Guinea (Port Moresby, Lae and Goroka).

Attitudes to punishment

Opinions on sentencing also present an important measure of citizens' reactions to crime which might be related to victimisation experiences, although only further analysis can show whether it holds true and for which types of crimes and sentences in developing countries. The question referred to five types of sentences considered most appropriate for a recidivist burglar (a man of 21 years of age who has stolen a colour TV) followed by a specification of the length of imprisonment should this sentence be chosen by the respondent.

[ Figure 29 ]

There is a high degree of agreement among the population in the developing world that the most appropriate sanction is imprisonment: more than 50% in all the regions, and even more than 70% in Sub-Saharan Africa and Asia. Citizens of Beijing and Manila in particular exhibit a punitive orientation. A fine is deemed the most appropriate sentence by 29% of the respondents from Bombay and 19% from Tunis, while community service is held an appropriate sentence by a bit less than half of the respondents in Rio de Janeiro. It is interesting to note that in Argentina and Brazil quite a few respondents favoured community service, although at the time of survey administration such a measure was not envisaged in the respective criminal justice systems. Nevertheless, punitativness prevails in the developing world.

Chapter 3
Towards a Comparison in an International Perspective

The preceding chapter presented selected key findings related only to the developing countries participating in the second round of the International Crime (Victim) Survey. An attempt towards truly global comparison is fraught with many difficulties particularly related to methods of data collection and sampling designs. In the industrialised countries the survey was carried out on a national sample using the 'Random Digit Dialing' technique and computer-assisted telephone interviewing. In the developing and Central and Eastern European countries the survey was carried out using face-to-face interviewing. In addition, in the developing countries as well as in Russia and Slovenia, the IC(V)S was carried out among the inhabitants of one or more major cities. The ensuing differences in methodology used preclude straightforward comparison. Therefore, in order to increase the comparability of data from the industrialised and Central and Eastern European countries on the one hand, with those from developing world city surveys on the other, special victimisation rates were calculated for the urban areas with more than 100,000 inhabitants for the nationally-based surveys. Obviously, the results should be read with caution. In addition, in this chapter the comparison is at a higher level of aggregation, namely, industrialised, Eastern and Central European and developing countries. No regional level comparison is provided for. As mentioned earlier, analysis on the international comparative level is underway.

Victimisation rates

[ Figure 30 ]

It is apparent that the developing countries exhibit higher victimisation rates for all selected types of crime, followed by urban areas in the Eastern and Central European countries. The smallest difference between the groups of countries relates to assault/threat, for which the developing and industrialised countries show similar risks. It should also be observed that the three groups show different rankings of the five types of crimes considered in Figure 1. While in all the groups the highest level is for personal theft (which represents at least one-third of the total victimisation in all the groups), burglary is the second highest in the developing and Eastern and Central European countries only. In the industrialised countries the second most frequent type of victimisation is assault/threat. On the other hand, in the developing countries, assault/threat is the least frequent type of victimisation observed. Robbery ranks last in the Eastern-Central European and industrialised countries and last but one in the developing countries, where risks of robbery almost level those for assault/threat.

Burglaries from homes were reported most often in the cities of the developing world, and particularly in Africa. Similarly, citizens of the developing world are more frequent victims of personal theft. Both for burglary and personal theft, the risk of victimisation is higher in the urban areas of Central and Eastern Europe than in the industrialised world.

In general, the risk of violence is more equally distributed across the world although robbery is more frequent in the developing world while the levels are similar in Central and Eastern Europe and the industrialised countries. Risk of assaults and threats are at a similar level for all three groups although it should be noted that it is a bit higher in the urban areas of the industrialised world in comparison to the Central and Eastern Europe.

Women in the developing world run the highest levels of risk for sexual incidents, although the effects of over- and under-reporting should not be disregarded.

Figure 31 presents a breakdown by type of crime in each group of countries.

[ Figure 31 ]

In all three groups more than 50% of the crimes are property-related. However, in the Eastern-Central European countries property-related crimes represent almost 60% of the total crime structure. The participation of sexual incidents in the total crime structure is equal in all three groups. It is interesting to note a comparatively higher proportion of violent crime in the industrialised group.

Vehicle-related crimes

[ Figure 32 ]

Concerning vehicle-related crimes, comparison is further complicated by drastically different ownership rates and structures. Ownership rates of cars are related to affluence, and there is a strong correlation between risk of car theft and levels of car ownership. It is apparent that the industrialised countries have a higher car related crime rate in general, and car theft rate in particular, indicating that high car ownership in the industrialised countries (on average above 80%) generates a higher probability of car theft. This is further associated with an outgoing lifestyle. However, in the developing and Eastern and Central European countries (with an ownership rate of less than 50%) the probability of car owners being victims is higher because of the unequal ownership structure and scarcity. Other factors will influence this relationship (e.g. quality of cars, usage of security devices, workings of the criminal justice apparatus, etc.). Stolen car recovery rates appear to indicate different usage of stolen cars: a high rate implies a prevalence of theft for the purpose of joy-riding, while few cars recovered may indicate further marketing of stolen cars. The latter, in turn, may also indicate a presence of organised crime involved in trafficking cars both in the domestic market and across the borders. On the one hand, this is an example of a response by organised crime to a market demand for cars, particularly in some Eastern and Central European, and some developing countries. On the other hand, it is also an indication of a need for better car protection systems and devices, more suitable car insurance policies, and improved police capability in vehicle recovery procedures.

Corruption and consumer fraud

[ Figure 33 ]

Corruption is an ubiquitous problem, although in the 1992/94 survey corruption was dealt with only in the developing and Eastern and Central European countries. There was a plight of corruption in the developing world in particular, in which it is one of the commonest forms of citizen victimisation. Consumers are cheated and badly treated particularly in Dar Es Salaam, Tunis, Georgia, former Czechoslovakia and Moscow. Both forms of victimisation appear to indicate a lack of consumer/client protection and the ways in which people go about, or are made to go about, satisfying their needs and rights. There is a large difference in consumer fraud rates between the developing world and the Eastern and Central European region on the one hand, and the industrialised countries on the other. As noted in the 1991 World Development Report, this can be explained by patterns of behaviour which adapt to market needs and influence transactions costs. 'Only after markets become established, when transactions become regular and competition increases, do traders have an incentive to establish and maintain their reputation (...), not only because sanctions are administered more efficiently, but also because a good reputation reduces transaction costs'.

Fear of crime

Fear for safety is generally seen as an important element of the social cost of crime, and crime prevention policies are geared not only towards reducing and controlling crime and crime-generating opportunities, but also anxiety and feelings of insecurity. Fear of burglary and street crime is on average higher in the developing and Central and Eastern European countries with the exception of Asia where both 'fears' are by far the lowest. 'In a regression analysis, level of urbanisation and burglary rates were both independently of each other strongly related to fear of burglary...Fear of crime, too, closely reflects actual exposure to violent crime...criminal victimizations substantially increase the awareness of crime risks among both victim and the public at large'.

Satisfaction with the police

The frequency of reporting to the police differs according to the perceived seriousness of the event which, in general, is linked to the type of crime as well as to many factors related to actual or expected police behaviour. Therefore, on average crimes are not reported because they are deemed 'not serious enough'. Frequently, in the majority of the developing and Central and Eastern European countries, it is a belief in or experience with the lack of effectiveness of the police which figures prominently. In general, people in the industrialised world report a much higher satisfaction with police performance than in the developing and Central and Eastern European countries. In those countries the majority of respondents were not satisfied with the police and showed a marked lack of confidence in law enforcement capacities, capabilities and even willingness to serve the community.

[ Figure 34 ]

One of the factors influencing the reporting rate for property crimes, such as burglary, is the presence or lack of insurance. The level of household insurance coverage is much higher in the industrialised countries than in the developing world (70% as opposed to 10-20%). 'At the aggregate level there is obviously a strong association between the extent of insurance coverage and reporting of burglaries to the police'.

The level of satisfaction with the police after reporting crime was much lower in the developing and Central and Eastern European countries as compared with the industrialised world. In the developing countries, as noted, almost one-third of victims who were dissatisfied with the police singled out lack of commitment on the part of the police ('didn't do enough') followed by unmet expectations as to the recovery of stolen property and/or bringing the offender to justice ('didn't find the offender'). In Eastern and Central Europe also, many victims complained that the police did not recover property, or were slow to arrive or did not find the offender. On the other hand, in the developed world, particularly with respect to property crimes, the outcomes of investigation, including the recovery of stolen property, is of less concern due to insurance coverage. In these cases the contact with the police and satisfaction is more administratively-oriented, such as efficient treatment and provision of documents supporting the insurance claim.

Attitudes to punishment

Opinions on sentencing also present an important measure of citizens' reactions to crime which might be related to victimisation experience, although only further analysis can show whether it holds true and for which types of crimes and sentences. Different patterns were identified: imprisonment is the most frequently chosen sentence in the developing and Eastern and Central European countries, with the exception of Poland and Ljubljana. In contrast with the stereotypical image of public demand for imprisonment, community service orders are seen in most industrialised countries as the most suitable punishment.

[ Figure 35 ]

The preference for imprisonment in developing countries has a lot to do with at least two sets of factors: first, on average the lesser availability of non-custodial sentencing options in most of the developing countries as well as difficulties in implementation following the conviction; second, support for imprisonment is higher in crime-ridden societies, particularly when no other available and practical solutions exist. However, certain factors related to cultural and socio-legal heritage also influence attitudes towards sentencing. For example, in the developed world the demand for imprisonment was much higher among 'anglophone' countries independently of personal victimisation experience or other crime-related factors. In any event, the experience in Finland - where community service orders were introduced on a larger scale -, or Brazil - where at the time of the survey public debate took place on the introduction of the community service order -, indicate that the public may become more supportive of alternatives to imprisonment if properly informed.

Policy implications

The results of the International Crime (Victim) Survey indicate that crime is a heavy burden on many people with, for instance, over one in five having experienced in the last year at least one incident of theft or damage to their property, or some form of aggressive behaviour. Levels of actual risk are far from negligible, whether or not these are cushioned by insurance premiums or social support networks.

At the same time, the international survey results help put local crime problems in perspective. In many Western countries, the public view is probably that crime is a 'national plague' for which local shortcomings are to be blamed. There may well be little awareness that other countries with different political and economic systems, and governments or, for instance, different family/educational infrastructures face similar problems. The survey data clearly dismiss the notion of high crime rates as unique to just one or two countries. With the exceptions of Japan and Switzerland, all industrialised countries suffer from an appreciable level of property and violent crime, particularly in the more urbanised areas. With the exception of one or two Asian and Central European cities, all the developing and Eastern and Central European countries also suffer from an appreciable level of property and violent crime. Victimisation by crime has become a universal experience or expectation for those living in large cities all over the world.

Within this global picture the developing urban world is in particular exposed to a number of social change processes which may further contribute to increases in victimisation, such as rapid urban/rural migration, refugees, population growth, decrease in job opportunity, problems with adequate health and educational services, changes in the family and neigbourhood structures, penetration of organised crime, etc. These challenges of a global and local nature require strategic but targeted responses. Yet their credibility and effects to a great extent depend on specific policy responses, including those in community-centred crime prevention. Particular attention should be given to proper information on crime risk, crime prevention and victim support. A rational approach to crime and crime prevention concerns both those officially in charge of criminal justice administration as well as citizens - actual and/or potential crime victims. At the same time, adequate efforts are needed to encourage citizens' responses to crime risks through participation in community crime prevention programmes, self-adoption of precautionary measures within the limits of accepted criminal policy and regulations, and increased co-operation with, and confidence in the law enforcement and other criminal justice agencies. On the other hand, criminal justice policy-makers are invited to promote programmes which are efficient, acceptable and credible to the community. An important part of these efforts regards the development of democratic, accountable and community-service oriented law enforcement which will pay special attention to prevention and respect for citizens and victims of crime. The promotion of special, and reliance on existing victim support structures and schemes, as well as recognition of victims' claims and rights in the criminal process is also of fundamental importance. In other words, serious commitment towards people's and community needs and rights is on the agenda. This requires more, but at the same time, less than investments in criminal justice equipment for crime and public order control. However, it should not be neglected that crime prevention, self-protection and tight control initiatives have their own limits, particularly in the developing world in which 'a demand-driven crime problem' prevails. Development brings about new opportunities for crime changing the crime structure and the crime/victimisation profile. But development potentially also brings new opportunities for effective crime prevention and equitable administration of criminal justice. This, however, depends on certain deliberate choices and strategies, including protection and respect for community and victims' needs and rights. Victim surveys can assist in this process.