Expressive Crimes in Post-Socialist States of , Latvia and Lithuania1 VAˆ NIA CECCATO Department of Urban Planning and Environment, Royal Institute of Technology, Stockholm, Sweden

Abstract

This article presents trends in availability and quality are dis- predict the variation of 2000’s expressive crimes in Estonia, cussed. While police official sta- expressive crime ratios than Latvia, and Lithuania from tistics show a significant rise in other indicators, such as land 1993 to 2000 and examines how rates of expressive crime in the use and economic covariates. demographic, socio-economic, Baltic countries during the 1990s Most of these covariates function land use, and institutional fac- (with the exception of homicide), in ways which are predicted by tors relate to their geography in victimization crime surveys indi- Western literature on crime geo- 2000. Geographical Information cate that there have been no graphy. System (GIS) and spatial regres- significant changes in crime sion models are employed in the levels and composition. Results key words: Assault, Crime geo- study, which make use of coun- also show that indicators of graphy, GIS, Homicide, Post- try regions as the unit of analy- regions’ social structure, such socialist countries, Regression sis. Issues concerning crime data as divorce rate, more strongly models, Trends, Vandalism

Introduction in these states during this period, or the The Baltic states of Estonia, Latvia, and changes that have taken place, and why Lithuania have experienced striking but (but see, for example, Butler 1992; not identical increases in the number of Aromaa 1998; Lehti 1998; Van Duyne reported offences during the last decade 2001; Rawlinson 2001; Ahven 2004). (Nordic Criminal Statistics 2003; These three Baltic countries represent Interpol 2003). The highest increase interesting case studies for scientific and was reported in Estonia, where the rate policy-related research into the geogra- increased from 245 reported offences per phy of crime (Fig 1). First, since the 10,000 inhabitants in 1993 to 421 in collapse of the Soviet Union, these states 2000, followed by Lithuania, where it have undergone a period of profound rose from 162 to 235 reported offences, political, economic, and social change. and Latvia, where it rose from 155 to These changes are expected to have 209 reported offences (Appendix 1). implications for the level and composi- However, relatively little is known about tion of offences as well as their geogra- the composition and geography of crime phies. Second, Estonia, Latvia, and Lithuania have recently joined the EU. 1This paper is part of the project ‘States in transition Historically they have been transit coun- and their geography of crime’, financed by The Bank of tries between East and West (Ulrich Sweden Tercentenary Foundation (Riksbankens Jubileumsfond, grant J2004-0142:1). A preliminary 1994); however, now they have become version of this article was presented in the 6th Annual border states on the critical eastern Conference of the European Society of Criminology, Tubingen, Germany, 2006. border of the EU. They are required to

2 Journal of Scandinavian Studies in Criminology and Crime Prevention ISSN 1404–3858 Vol 9, pp 2–30, 2008

DOI: 10.1080/14043850701610428 # Taylor & Francis ceccato: expressive crimes in estonia, latvia and lithuania

Figure 1. The study area: Estonia, Latvia, and Lithuania. adopt a rigorous approach towards serve to vent rage, anger, or frustration. maintaining border security. EU enlarge- In this paper, expressive crime refers to ment increased the need for knowledge homicide, assault, and vandalism.2 that provides the basis for policies to Moreover, expressive offences are asso- deal with crime in Europe as a whole ciated with stress and social disorganiza- when the three Baltic states became part tion. In Russia and Eastern European of that larger picture. countries, economic and political transi- This paper has two objectives. The tion has been associated with uncer- first is to present crime trends in Estonia, tainty and social instability which has Latvia and Lithuania covering the years following their independence and transi- 2As in other previous studies (e.g. Western et al. tion from a planned to a market 2003:47; Perry 2003:287) vandalism is regarded as an economy (1993 to 2000). The analysis expressive crime because it is a form of expression. Vandalism reflects youths’ ‘expressive emotional out- will focus on expressive crimes because lets’ in a social situation when they experience the rates of expressive crime have frustration (Mays 1970:53), expressed by acts that serve to vent rage and anger. As Sampson and Raudensbush remained consistently high in these (1999:609) suggest, ‘vandalism does not directly cause other more serious crime but creates a neighbourhood countries (see Appendix 1). Expressive environment where they are more likely to occur’, crimes are defined as criminal acts that indicating they may might share the same geography.

Journal of Scandinavian Studies in Criminology and Crime Prevention 3 ceccato: expressive crimes in estonia, latvia and lithuania generated grounds for violence. Recent that has increased the availability and studies in Russia and Eastern Europe validity of social, economic and vital suggest that social stress and disorgani- statistics data for scientific research. The zation during transition from a planned study also contributes to the issues of to a market economy are related to data availability and quality by gathering increases in suicide, homicide, and over- information through a survey of experts all mortality (Leon and Shkolnikov 1998; in each country who work directly with Gavrilova et al. 2000; Pridemore and crime statistics. Spivak 2003). Official police data from The structure of this article is as the Baltic countries are compared with follows. It begins by discussing issues on those of Europe, as well as other data data availability and quality in the Baltic sources, such as victimization surveys. countries, particularly crime data. The second objective is to assess how General trends in expressive crimes are demographic, socio-economic, and insti- presented as a background for the con- tutional factors relate to the geography ceptual framework for the case studies of of expressive offences in the Baltic Estonia, Latvia, and Lithuania. The region in 2000. Geographical Informa- relationship between crime and the pro- tion System (GIS) and spatial regression cess of change is discussed and hypothe- models are used, with ‘country regions’3 ses of study are suggested. This is serving as the unit of analysis. Since they followed by the modelling section, where constitute three independent countries, results are discussed. The article con- national differences are also tested in the cludes with implications of the findings analysis. The novelty of this article comes and directions for future work. from its inclusion of the spatial dimen- sion of crime, which is often neglected in Crime data in the Baltic countries the recent literature devoted to the post- No recommendations have yet been socialist countries. Until recently, spatial made to unify data collection, processing crime analysis in these nations in transi- methods, or analytical procedures for tion was rare simply because data were European crime statistics. Each European not systematically available for all country uses its own system of defini- regions, or data quality was still a major tions, with distinct rules for collecting limiting factor. As Pridemore (2000) data (Gruszczynska and Gruszczynski suggests, one benefit of this transition 2004).4 For a discussion of the common has been an improvement in transparency problems with police-recorded data, see the European Sourcebook of Crime and 3They are polygons from SABE 2004—Seamless Criminal Justice Statistics (1996, 2003). Administrative Boundaries of Europe (Eurogeo- graphics 2005), which is a pan-European data set Another common problem with crime containing the geometry and semantics of the adminis- data from this source is under-reporting. trative hierarchies of 36 European countries. The average population size in 2000 for the regions in the three Baltic countries was just under 70,000. These 4Particularly for homicides, mortality data are com- regions are administrative units at county level, such as monly considered to provide a basis for comparison with in Estonia, and in some cases municipality adminis- police crime records (e.g. Stickley and Ma¨kinen 2005). trative regions, such as in Lithuania. Although they Since they often show the same trends, including for the differ in administrative level, they are territorially very Baltic countries (see e.g. Ahven 2002), we are going to similar in size. focus this analysis only on police official statistics.

4 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

According to the International Crime The following paragraphs describe Victimisation Survey (ICVS) (Del Frate what these experts reported about and Van Kesteren 2004) the percentage of how and when different factors or events crime reported to the police was con- may have affected crime data quality. siderably lower in Central and Eastern Experts from each country were also European countries than in Western invited to modify crime definitions of Europe throughout the 1990s. the selected expressive offences (initially In the particular case of the Baltic based on the European Sourcebook of countries, crime records were influenced Crime and Criminal Justice Statistics) by political and administrative changes according to the local penal codes. This following independence and are still paper focuses on expressive crime, but affected by changes in criminal codes the discussion about crime data quality and police practices, including corrup- encompasses other types of offences as tion. Most of the major systematic well. changes are well known (e.g. changes In Estonia, crime data are gathered by in penal codes and recording systems), central criminal police. According to but the difficulty lies in estimating how experts, there have been two major minor (perhaps local) changes affected changes in the way crime has been levels of recording. According to defined following independence. In Gruszczynska (2004), in many Central 1992 the Estonian criminal code was and Eastern European countries interna- adopted, abolishing the differentiation tional projects play an important role in between crimes against state property improving their statistical standards. and crimes against private property, a Although this article recognizes that hall-mark of the Soviet criminal code. the quality of crime data has improved The Soviet code also included some since independence, little is documented crimes which were in essence political about data quality in these countries crimes—all such articles were abolished (e.g. differences arising from the variable with the new code. The second major structure of police authorities, from change took place in 2002. After 1st regional ethnic language differences, September, a new penal code replaced or from technological barriers which the criminal code and part of the would complicate the implementation of new recording systems). To provide a administrative code was completely dis- better picture of data quality, a survey solved and its provisions transferred to was performed with a small group of or replaced by other laws. Repeated use professionals dealing with crime data.5 of illicit drugs or possession of a small amount of illicit drugs for personal use 5The questionnaire used in this survey is available on was decriminalized and reclassified as request. Fourteen experts answered the questionnaire, misdemeanours. However, the total which includes questions on availability and quality of data from the early 1990s to the present from co- number of drug offences regarded as ordinate (x,y) to the national level, changes in penal criminal (e.g. drug possession with code, police organization, processes of data gathering and systematization, and boundary/administrative intent to supply, drug trafficking) and changes in police units over time. A list of the experts misdemeanours (drug abuse or posses- who participated in this survey is shown at http:// www.infra.kth.se/sp/Research/Transition/analytic1.htm sion of a small amount for personal

Journal of Scandinavian Studies in Criminology and Crime Prevention 5 ceccato: expressive crimes in estonia, latvia and lithuania use) is comparable to those under the The structure and content of this crim- previous code. Changes in the definition inal code differed from the Soviet of assault from the criminal to the penal criminal code (1961–1999) in supporting codes make data related to assault historical traditions of independent incomparable between the two. General Latvia (1918–1940) and the state’s poli- statistical comparisons are also possible tical ambitions to join the EU. After for thefts (including subtypes by independence, new criminogenic condi- objects), although there were some tions required new crime codes, such as changes in wording. The recorded rates smuggling, while some offences no of several offences have been affected by longer qualified as crimes, for instance these changes. A significant increase in buying and selling goods in order to earn the number of drug offences reflects the money (making profit). According to the police’s increased capacity to tackle drug experts, the major changes in recorded crime as well as an actual increase in crime occurred when the Latvian system drug use and trafficking. Driving under of recording offences replaced the the influence was criminalized in 1999, Russian one. Experts believe that crime increasing the number of recorded data quality has improved since the offences in the following years. The implementation of this ‘new system’ criminal code for robbery and a subsec- (when paper records were replaced by tion for aggravated unconcealed theft a computerized system). were ‘related’, and such crimes are now In Lithuania, crime data are system- described in the penal code as robbery. atized by the Ministry of the Interior. Therefore the total number of robberies This is a centralized registration system increased rapidly after implementation which has been responsible for format- of the last penal code in 2002. ting crime data, controlling the quality, In Latvia, the Information Centre of and making data public since 1971. the Ministry of the Interior is respon- These records contain data on the sible for systematizing crime data. criminal offence, investigation results, According to experts, the most signifi- charges, investigation procedure and, cant changes took place between 1994 since 2003, the victim. Crime recording and 1996. These changes focused on had already begun to shift into a digital criminal procedures, such as which format in the 1980s, when Lithuania was authorities would deliver indictments still part of the Soviet Union. All or prosecution, rather than the recording criminal offences that took place in of crime data, which was only minimally Lithuania were recorded on ‘cards’ and affected. In 1999 the new Criminal Law later sent to the Ministry of the Interior. came into force but started to be applied Crime recording became computerized from 2000 onwards. According to the as early as 1994 but was not nationalized new Criminal Law, criminal offences until 2003, before which all police have two types: criminal violations commissariats had unique data systems. (warranting no more than two years’ On May 1st, 2003, a new penal code deprivation of liberty) and crimes (less changed the method of recording police serious, serious, and especially serious). crime data in Lithuania, which partially

6 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania explains the increase in the number Estonia, Latvia, and Lithuania had of registered crimes in international significantly higher homicide rates than statistics. In most cases, this was due to other countries in Europe throughout changes in crime definitions or in the the 1990s (Fig 2). The lowest rates were status of certain activities as criminal found in Sweden, Denmark, Norway, (e.g. smuggling of goods, misdemea- and the Netherlands. According to the nour). Theft was most affected; before European Sourcebook (2003), homicide 2003 minor thefts were regarded as rates vary significantly between coun- administrative offences but now quali- tries, even when attempted homicide is fied as criminal offences. Moreover, excluded. Other variations in definitions after 2000 there were several territorial may influence homicide rates but do not, changes in Lithuania (e.g. number and by themselves, explain these differences boundaries of the municipalities) that (see Appendix 1). However, in the last might have affected the recording of 50 years Estonia, Latvia, and Lithuania crime data at a municipal level. had relatively low homicide rates com- pared to Russia, for instance (Stickley Trends in expressive crimes in the Baltic and Ma¨kinen 2005). countries Among the Baltic countries, Estonia The analysis will concentrate mostly on has historically shown higher homicide police-recorded data from 1993 to 2000 rates than its Baltic neighbours (Lehti because this represents a period of 2001). In the particular case of Estonia, relative stability following the countries’ Ahven (2004) indicates that a factor independence. Before that, it is difficult contributing to the rise in homicides in to ascertain the quality of the data, the beginning of the 1990s was that the which was often filtered by the autho- availability of alcohol increased mark- rities because a rise in crime was edly after 1992. While conflicts in the interpreted as a threat to the former political system (see e.g. Butler 1992; criminal underworld are also considered Lehti 2001; Gruszczynska 2004). After a leading cause of the increase in 2000, changes in the penal codes in these homicide in the early 1990s, later countries have affected the way offences analyses have shown that such conflicts are recorded. Moreover, there have been account for less than 20% of homicides. territorial changes in the Baltic countries Nearly two-thirds of homicides were that may have affected the way crime is committed by acquaintances or family recorded (this is particularly true for members after drinking together. Since Lithuania). 1995, homicides have decreased consid- In order to gain a general under- erably, reaching the same level as at the standing of crime trends, police official beginning of the 1990s. statistics in the Baltic countries are first Only homicides, out of three selected compared with trends in police data expressive offences, decreased during the from Western European countries (see 1990s (more than 30%). Reduction in Appendix 1), then crime trends based on homicide rates was also witnessed in police data are compared with trends in other countries such as Germany and data for victimization surveys. Finland, and in Eastern Europe, in

Journal of Scandinavian Studies in Criminology and Crime Prevention 7 ceccato: expressive crimes in estonia, latvia and lithuania

2.5

Estonia Latvia 2 Lithuania Czech Republic Poland Slovenia 1.5 Norway Sweden Denmark

1 Finland England & Wales Germany France

Homicide rates per 10,000 inhabitants 0.5 Italy Netherlands Portugal

0 1993 1994 1995 1996 1997 1998 1999 2000

Figure 2. Homicide rates per 10,000 inhabitants, 1993–2000, in relation to other EU countries.

Slovenia and Bulgaria (Fig 3). However, Latvia and Lithuania, including vandal- levels of aggression as a whole did not ism, pick-pocketing, and drug-related decline in the Baltic states, indicating crimes, resulting in significant increases that violence in the 1990s may have in recorded rates. become less lethal. According to interviewed experts, as Assault, for instance, increased more overall living standards improved since than 50% according to police statistics the mid 1990s, crime levels have become (Fig 3), particularly in Lithuania and more similar to the ones found in Estonia. A similar trend was also found Western European countries. They in Western European countries6 such as believe that official figures for the early Belgium, Denmark, Finland, Norway, 1990s did not reflect actual crime, at and France. Vandalism, which can also least in comparison with Western or be considered a form of aggression, has Nordic countries, due to differences in increased more than 100% in the Baltic legislation, statistical procedures and the countries since 1993. It is also worth police capability and effectiveness to noting that reporting improved in the detect some types of crimes. Experts 1990s for certain crime types in Estonia, suggest that the police (and other criminal justice institutions) have become more efficient in targeting cer- 6According to the definition used by Gruszczynska and Gruszczynski (2004), this includes old European Union tain types of crimes. For instance, co- members, plus Norway and Switzerland. operation between private security firms

8 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

Figure 3. Expressive crime rates per 10,000 inhabitants in Estonia, Latvia, and Lithuania, 1993–2000. and the police intensified and improved Victimisation Survey (ICVS) and the their capacity in catching shoplifters. Living Conditions Survey (LCS).7 Some changes in criminal legislation and Findings also indicate that the level and statistical procedures may also have structure of crime have remained largely influenced official trends (e.g. increasing the same during this period of time. The registration of physical abuse). Experts LCS from 1994 and 1999 showed that tend to suggest that an increase in reports of victimization by violent official statistics in the 1990s does not attacks and threats have decreased in necessarily mean actual increase in Latvia and Lithuania since 1994, but crime—a statement that is supported showed almost no change in Estonia. by the results from victimization surveys These divergent trends between data in the three Baltic countries. from the police and victimization sur- Whilst official police figures suggest veys make the interpretation of crime crime rates increased several-fold in the trends in the Baltic countries a difficult Baltic states between 1995 and 2000, task. victimization surveys indicate there have been no essential changes. Ahven (2002) 7The International Crime Victimisation Survey (ICVS) compared trends in police crime data was carried out in Estonia in 1993, 1995 and 2000; in from the second half of the 1990s with Latvia (Riga region) in 1996, 1998 and 2000; and in Lithuania (Vilnius) in 1997 and 2000. The Living survey data for the Baltic countries from Conditions Survey (LCS) was performed in all Baltic two sources: the International Crime countries in 1994 and 1999.

Journal of Scandinavian Studies in Criminology and Crime Prevention 9 ceccato: expressive crimes in estonia, latvia and lithuania

A possible cause of this is the growth violence are related to a certain extent to of crime reporting to police in the Baltic differences in the population’s ethnic countries. In Finland, for instance, the composition. According to police statis- long-term growth of the crime reporting tics, the violent criminality of the non- rate was consistent with the growth of Estonian (mainly Russian) population has fear, the increase of the average age of been higher than the Estonian population the population, the on-going urbaniza- for several decades. People of non- tion and the gradual improvement of Estonian origin, who have made up about the availability of police services one-third of the population since the (Heiskanen et al. 2004:24). In the light 1970s, have committed about two-thirds of previous studies in Great Britain, the of homicides (regardless of great fluctua- United States and elsewhere, this trend tions in general homicide levels in the can also be associated with an increase 1980s). According to police data, in 2005, in victimization-related inequality (see for people without Estonian citizenship (18% instance Trickett et al. 1995; Levitt 1999; of the total population) committed 43% Young and Mathews 2003; Nilsson and of solved homicides. Although this ten- Estrada 2006). In Sweden, for instance, dency also appears in the homicide survey Nilsson and Estrada (2006) suggest that (Lehti 1998) and was also confirmed by the reason behind the overall stability in interviewed experts, comparisons should victimization is the increasing polariza- be interpreted with caution. One reason is tion between different groups of society. that the category ‘non-Estonian origin’ Richer groups had their exposure to does not necessarily mean ‘foreign-born violence stabilized since the mid 1980s, population’ as in the case of most Western whilst the proportion of those victimized European countries. Moreover, this pat- has become significantly greater among tern of ethnic minorities and expressive the poor. For the Baltic countries, how- crimes does not repeat itself in Lithuania. ever, inequality in victimization cannot be In 2000, 28% of offenders involved in tested since socio-economic data from homicides were ethnic minorities (the victimization surveys are too limited to proportion of ethnic minorities in carry out meaningful analysis, or the Lithuania was 17% of the whole popula- sample size of respondents varies over tion in the same year). Data were not time (some were applied to the capital available to confirm such a trend in cities only, whilst others covered the Latvia. whole country). Regional differences in violent crimes in Latvia are less abrupt than those Regional patterns of crime: Estonia, found in Estonia (Lehti 2001). In districts Latvia and Lithuania with high unemployment rates, such as In Estonia, the level of expressive crimes is Rezekne and Preili, expressive crimes, highest in and in the cities of Ida- such as assault and vandalism, tend to Viru county. For an overview of crime be higher than the national average. As levels in Estonia, see Saar et al. (2003) and in Estonia and Latvia, crime is more Ahven (2004), and on violent crimes see wide-spread in urban than in rural areas Lehti (2001). Regional differences in in Lithuania. The three largest urban

10 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania areas (Vilnius in the south-east of the available for the calculation of crime country, Kaunas in the centre and rates. Total population was the chosen Klaipeda at the Baltic Sea coast) are denominator for calculating SORs for all densely populated, relatively affluent, expressive offences. Using resident popu- and experience some of the highest rates lation as a denominator is not problem- of reported offences in the country. The free since there are regions, such as the Baltic countries’ location also accounts capital cities and other holiday resorts, for organized crime groups involved in that receive a significant flow of tempor- trafficking different goods, including ary population during certain periods of illegal weapons, stolen vehicles, and year, mostly the summer. Since crime contraband products such as cigarettes, ratios are calculated based on resident clothes, furniture, counterfeit money, population, the risk of victimization and illegal immigrants (Europol 2004), might not be accurate for these regions. which might also have an effect on the Fig 4 illustrates standardized ratios for level of expressive crimes. homicides and vandalism. Shifts have occurred in the geography of both homi- Standardized offence ratios were cal- cides and vandalism between 1993 and culated in order to identify regions that 2000. The regions in dark grey reported run a relatively high risk for certain more offences than expected, while the expressive crimes, taking into account regions in light grey reported fewer the overall distribution of crime in the offences than expected, taking into con- Baltic countries. Ratio is a useful way of sideration the whole Baltic region. For representing data for a set of regions homicide, note that Vilnius is the only where the areas differ in size, as in this capital city where the observed number of case (absolute values would tend to homicides was lower than expected both over-emphasize large areal units), and in 1993 and 2000. where it is necessary to allow for differences in population characteristics Crime and processes of change in the (on the advantages of using ratios Baltic countries instead of rates, see Haining 2003). The Although the heightened levels of crim- standardized offence ratio (SOR) for inality in post-socialist countries have region i is the ratio between the observed historic roots, as suggested by Lotspeich number of offences and the expected (1995), there are reasons to believe that number of offences. In this analysis, an recent transformations in the Baltic average offence rate was obtained by countries have also impacted their crim- dividing the total number of offences by inogenic conditions. The international the total size of the chosen denominator. literature has a long tradition of search- For each area i, this average rate is ing for links between changes in multiplied by the size of the chosen countries’ socio-economic structural denominator in area i to yield E(i). conditions and crime. Durkheim (1897) Wikstro¨m (1991) pointed out the diffi- was the first to argue that rapid social culty of defining appropriate denomina- change creates anomie, which can have a tors and suggested a list composed of negative impact on society and may lead best denominators and those which are to crime. The link between social change

Journal of Scandinavian Studies in Criminology and Crime Prevention 11 ceccato: expressive crimes in estonia, latvia and lithuania

Figure 4(a). Standardized homicide ratios for Estonia, Latvia and Lithuania. Figure 4(b). Standardized vandalism ratios for Estonia, Latvia and Lithuania.

12 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania and crime occurs because rapid transfor- Although these macro-changes are pre- mations (here associated with the indus- sented separately, they should be trial society) produce a chronic state of regarded as simultaneous, creating par- deregulation, in which valued goals ticular crime conditions both at national become discounted and society fails to and regional levels. place normative limits on peoples’ desires. For American theorists, such as Demography Merton (1967), anomie is caused by a Estonia, Latvia and Lithuania have fol- lack of adequate means to fulfil society’s lowed the same trends as Russia and other culturally sanctioned goals (Orru` 1987) Central and Eastern European countries, conceptualized as ‘the American Dream’. experiencing low fertility, high mortality, Regardless of these theoretical differ- and high divorce rates, as well as emigra- ences, ‘normlessness’ translates into a tion significant enough to reduce the lack of or weakening of social controls population (see for example Philipov which are fundamental for the function- and Dorbritz 2003). Since gaining inde- ing of social institutions such as family, pendence, these countries have made an education, and polity, and therefore may effort to reintegrate their ethnic minorities be associated with changes in crime levels but challenges remain (see e.g. Vetik in societies in transition. Messner and 1993). Differences in life-style and long- Rosenfeld (1994, 1997) suggest that in a term socio-economic marginalization dominant capitalist society, social institu- have generated greater crime rates tions tend to be devalued in comparison amongst certain ethnic groups (e.g. to economic institutions (what has been (Lehti 2001) and called ‘institutional anomie’) and lose Slavic minorities in Latvia (Aasland and their power to positively influence crime Flotten 2001)). These demographic trends rates. In other words, the power of more had a negative impact on the population socially oriented institutions that have a of the Baltic countries, their internal potential control effect on crime becomes spatial distribution and certainly the reduced. However, there is empirical socio-economic features associated with evidence that negative market forces can this development. Violent crime is be moderated by social institutions in expected to be higher in areas character- countries that have strong social capital ized by population decrease and social (e.g. Chamlin and Cochran 1995) or instability, which does not necessarily maintain their pro-social systems (e.g. mean rural areas only. In Russia, for welfare) despite the market’s dominance instance, trends in fertility, mortality, and (e.g. Savolainen 2000). migration have been associated with In this section a set of criminogenic anomic social conditions and crime (e.g. features are discussed in relation to Kim and Pridemore 2005). Therefore, it is conditions in the Baltic countries in the expected that in the Baltic countries, 1990s. The focus is on the impact of anomic feelings will generate violence these processes of change on expressive and disorder where population decay crime and how they may have helped to and signs of deprivation and social delineate crime ratios in the year 2000. isolation are evident.

Journal of Scandinavian Studies in Criminology and Crime Prevention 13 ceccato: expressive crimes in estonia, latvia and lithuania

Socio-economic and welfare (Pampel and Gartner 1995; Kim and characteristics Pridemore 2005). This evidence suggests Another factor stimulating criminality is that crime levels would be lower in areas the deterioration of economic conditions. with high expenditure in social care, In the early 1990s, complex shifts in the despite negative socio-economic develop- type and level of economic activity caused ment (e.g. high unemployment rate). In significant drops in economic production regions where there are negative struc- and an increase in unemployment. In tural developments, crime levels are economies such as the former Soviet moderated by ‘collectivistic’ approaches republics, the process of decline in the to supplying public resources. manufacturing industry cannot be disas- sociated from the process of privatization Social cohesion of state companies. At the same time, the The literature on social cohesion has information economy has emerged and long suggested the importance of social had profound impact in more prosperous institutions in moderating the negative areas. Despite relatively high GDP per effects of structural problems in society, capita in these countries, unemployment including crime. Chamlin and Cochran remains a challenge (World Bank 2005a, (1995) show evidence of the suppressing 2005b). The market economy stimulates effect of social cohesion on poverty as a patterns of consumption that are not motivation for property crime in the US. attained by all segments of society. This Findings revealed that a lower divorce is expected to cause frustration arising rate, higher voter turn-out, and greater from the discrepancy between the goals of church membership reduced the effect of the free market economy and the ‘real’ poverty on property crimes. Divorce capacity to obtain them. The chaotic rate, an indicator of family structure, economic conditions that characterize has been suggested as a strong predictor the transition process engender criminal of offending. Sampson (1986) studied behaviour in a number of ways, particu- American cities with populations above larly in the domain of expressive crimes. 100,000 and found that the lower a city’s Income support via welfare, such as divorce rate, the lower its crime rate. pensions and allowances, has decreased One of the mechanisms that links significantly. Some areas became more broken families to offending is the market-oriented, privatizing many basic increase in poverty, particularly follow- services, while others struggled to main- ing a divorce (Corcoran and Chaudry tain their welfare system. This impacted 1997). Compared to other EU member most significantly on those regions char- states, divorce rates are relatively high in acterized by high unemployment rates the Baltic countries (Eurostat 2003). It is and poor government support (e.g. few expected that regions with higher measures to improve labour force skills or divorce rates will experience higher low municipal expenditure on social levels of expressive crimes. care). There is evidence that social Low voter turn-out is often regarded institutions moderate negative structural as an indication of weak social cohesion effects and therefore help reduce crime (e.g. Putnam 1995). In industrialized

14 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania democracies, low participation in elec- was a reduction in resources for govern- tions is linked to mistrust in society, ment activities, including law enforce- economic inequality, and a ‘weakening ment. These and a number of other of the norms governing behaviour’ aspects of post-socialist countries, such (Sampson et al. 1997:41). Declining as a region’s location and geography (e.g. participation in recent elections has differences in land use), degree of urbanity characterized newly democratic Eastern and the country’s relative position to European countries, even when com- organized crime, create conditions con- pared to participation rates in the West. ducive to crime (see e.g. Christie et al. The participation in parliamentary elec- 1965; Ulrich 1994; Ceccato and Haining tions in the Baltic countries dropped 2004). All these aspects are considered significantly just after independence into intervening factors affecting crime levels the late 1990s. In Estonia, for instance, in the Baltic countries and regarded as in regions such as Ida-Viru, participation control variables. in parliament elections was as high as 70.1% in 1992, dropped to 67.2% in Modelling expressive crimes 1995, and reached 57.9% in 1999. Low participation in elections is interpreted Dependent variables here as a sign of weakening social bonds, The dependent variables in this study are a lack of engagement in the newly the standardized offence ratios presented implemented democratic process, and above (Trends in expressive crimes in general mistrust in society. Crime the Baltic countries), which create a becomes an option when ‘social bonds’ cross-sectional measure of relative risk are weak and one’s connection to society for homicide, assault and vandalism. fails. The lower a region’s voter turn-out Crime data from the year 2000 were is, the higher its crime rate. chosen for two reasons. First, changes in penal codes in Estonia, Latvia and Other factors Lithuania make any comparison of An important characteristic of post- crime levels after 2000 very difficult. socialist countries is the changing nature Second, Baltic countries have experi- of law enforcement (e.g. penal code), enced an increase in total crimes, and police practices and organization, and, looking at crime levels in the year 2000 perhaps more importantly, the ‘inability provides a reasonable time frame for of criminal systems to adjust rapidly’ to assessing how demographic, land use, imposed transformations. As Lotspeich economic and social transformations (1995) indicates, police agencies and might have affected crime since indepen- courts faced a new set of challenges in dence in the early 1990s. the 1990s accommodating new laws and The set of standardized offence ratio actions to protect private property. This values showed a highly skewed distribu- state of confusion made criminal activity tion. The raw ratios were transformed more attractive by reducing the like- using the square root or reciprocal trans- lihood of being caught and convicted. formation to produce a data set more Another challenge these countries faced nearly normal. The first column in Table I

Journal of Scandinavian Studies in Criminology and Crime Prevention 15 ceccato: expressive crimes in estonia, latvia and lithuania indicates which transformation has pro- whether countries are statistically differ- duced the best result and was therefore ent, after controlling for country-specific used as dependent variable. In order to test differences in the demographic, socio- for spatial autocorrelation on the resi- economic composition of the population. duals, a row-standardized binary weight This kind of model has slopes that vary matrix (W) was used that comprised non- according to the country (Lithuania was zeroentrieswhereiandjrefertoareasthat used as a base country). The inclusion of are adjacent. Only contiguous regions country dummies significantly increased were used in the analysis, which excluded, the number of covariates in the models, for instance, the Estonian islands. This which made modelling virtually impos- was done in order to build a weight matrix sible due to high correlation between that allowed the testing of spatial auto- some of the covariates and their dum- correlation on residuals and go further to mies. The strategy was then to run other models (e.g. spatial error, spatial models by each covariate to identify the lag), if necessary (as suggested by Kim and most important covariates from each Pridemore 2005:93). The regression analy- group (demography, socio-economic sis was implemented using GeoDa 0.9.5-1 and welfare characteristics, social cohe- (Anselin 2003) since the software has sion and other factors). The next step was regression modelling capabilities with a then to try simultaneously the most range of diagnostics that are not available important covariates from each measure in standard statistical packages (e.g. as well as its country dummies (say, from Moran’s I test of spatial autocorrelation Welfare, number of hospital beds per on residuals). inhabitant and its country dummies were chosen). This procedure has proven successful for all offences, allowing a Independent variables pre-selection of covariates for the final These models include individual variables modelling runs. Nine variables (high- with data from 2000 providing a cross- lighted in Appendix 2) were used in the sectional picture of these covariates.8 The final models. list of independent variables to be tested in the model encompasses some of the key correlates of crime rates suggested Results and discussion by the current literature (Appendix 2). Table I summarizes the findings of Country dummies were added to test regression models, including the signi- ficant variables. Despite performing 8Measures of ‘change’ from 1993 to 2000 were tested as transformations to try to correct skew- covariates (indices of economic, social and welfare ness of the dependent variable, non- change), but the results were poor in explaining 2000’s expressive crimes ratios. Although some of these indices normality of the residuals was still a were significant in predicting assault and vandalism, problem in two out of three of the models none of them were significant in explaining variations of ratios of homicides. Social change described only 10% (homicide and assault). For assault, the of the variation in assault ratios. For vandalism, the Ordinary Least Squares regression model indices of social and economic change were significant and explained 28% of 2000’s vandalism ratios. diagnostics revealed significant spatial Problems of non-normality, heteroskesdasticity and autocorrelation on the residuals. Spatial autocorrelation on residuals were present in these models. error models were then fitted to address

16 Journal of Scandinavian Studies in Criminology and Crime Prevention ora fSadnva tde nCiiooyadCiePrevention Crime and Criminology in Studies Scandinavian of Journal

Table I. Results of the regression analysis: Y5standardized offence ratios 2000.

Significant variables Regression diagnostics

Homicide OLS model Multicollinearity condition number 8.16 Reciprocal 0.07420.36DivorceRateLatv+0.009NonNativePopLatv+0.06YoungMalePopLatv Test of normality of errors Jarque-Bera Value 315.97; p50.00 (2.49) (25.14) (3.36) (4.60) Test of heteroskedasticity White Value 5.64; p50.77 t-values in brackets Test of spatial dependence of errors Lagrange Multiplier Value 0.09; p50.92 R2X100523.2% All covariates are significant at the 1% level

Assault OLS model Multicollinearity condition number 16.89 lithuania and latvia estonia, in crimes expressive ceccato: Square root 9.6120.0006PopDens+1.29DivorceRate20.051VoterTurnout Test of normality of errors Jarque-Bera Value 20.78; p50.00 (5.88) (21.78) (3.70) (22.59) Test of heteroskedasticity White Value 14.05; p50.12 t-values in brackets Test of spatial dependence of errors Lagrange Multiplier Value 6.19; p50.01 R26100524.7% All covariates are significant at the 1% level

Spatial error modela Test of heteroskedasticity White Value 9.25; p50.03 8.9820.0006PopDens+1.41DivorceRate20.047VoterTurnout+0.26l Test on spatial error dependence and non-normality test are not (5.26) (21.84) (4.40) (22.04) (2.18) available in GeoDa for this model z-values in brackets

R26100529.3% DivorceRate is significant at the 1% level, Voterturnout is at the 5%, and PopDens is at 10% 17 ceccato: expressive crimes in estonia, latvia and lithuania

these problems, but heteroskedasticity still remained a problem.

Value 3.05; For all expressive crimes, divorce rate 0.32

5 is significant in all models (Fig 5). p Divorce rates had a strong increase 0.34 5

p between 1990 and 2000 in the Baltic countries: in Estonia they rose from 49 Value 2.26;

Lagrange Multiplier per hundred new marriages to 77 (the

Value 5.70; highest within the EU), in Latvia they rose from 46 to 67, and in Lithuania Jarque-Bera White from 35 to 64. The highest divorce rates are in large cities (e.g. Tallinn, Tartu), among ethnic Russians, and in ethnically mixed families (for more details, see Philipov and Dorbritz 2003). There are reasons to believe that in the Baltic 0.10 5

Multicollinearity condition number 7.46 p countries social changes caused by the transition from a planned to a market economy have affected key social insti- tutions, such as the family. The pre- viously stated hypothesis that regions with a high proportion of juveniles from broken families would have higher rates of crime and acts of violence can be corroborated by these findings. Findings show that high ratios of expressive crimes and more motivated offenders are concentrated in regions with high divorce rates. However, for Latvia, although divorce rate is significant, it is

5.06)negatively related Test of heteroskedasticity to homicide rates. 2 Homicide ratios are relatively high in certain less urbanized regions, such as 0.02HospitalBedsLatv Test of normality of errors

2 Rezekne, Ludza, and Kraslava districts. In Latvia, high homicide rates are also related to a high proportion of young 42.5%

5 males (e.g. Daugavpils district, Jelgava 1.60DivorceRate 100

+ city) and a high proportion of non- 6 2 -values in brackets Test of spatial dependence of errors Significant variablesOLS model 4.86 Regression diagnostics All covariates are significant at the 1% level (6.57)t (6.08)R ( native population (e.g. in Riga and Rezekne). Young males are regarded both as potential offenders and victims for violent crimes. Regarding non-native Table I. Continued. Spatial lag model was tested but showed poorer performance than the error model. populations, differences in life-styles and a Vandalism Square root

18 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

Divorce rate (2000) 4.2 to 5.2 3.2 to 4.2 2.2 to 3.2 1.2 to 2.2

Figure 5. Divorce rates in the Baltic countries, 2000. long-term socio-economic marginaliza- culture, and the state of civil society (see tion of minority ethnic groups have e.g. Lewis 1997; Berglund et al. 2001). historically generated particular propen- Assaults are also concentrated in regions sities to offending, such as amongst with low population density but rela- Slavic minorities in Latvia (see e.g. tively high divorce rates (in Estonia e.g. Aasland and Flotten 2001). Pa¨rnu district, in Lithuania e.g. Zarasai As previously hypothesized, more district). The negative sign for popula- assaults are found in areas with low tion density in the model of assaults as participation in elections. As in Western an indicator of urban areas is surprising, countries, low voter turn-out is an since the literature showed plenty of indicator of weak social cohesion. In evidence that crime is more wide-spread post-socialist countries, low participa- in urban than in rural areas (Christie et al. tion in elections is related to the 1965) and tends to increase with the size maturity of the voting process, but as of the urban area. Ceccato (2006) found suggested by Bernhagen and Marsh that crime levels in Lithuanian regions (2007) it involves a diverse array of were highly determined by ‘urban areas’, combined explanatory factors, such as both directly by the variable ‘urban areas’ political resources directed to voting and and by other covariates that correlated election processes, deprivation, political with it (e.g. number of hospital beds per

Journal of Scandinavian Studies in Criminology and Crime Prevention 19 ceccato: expressive crimes in estonia, latvia and lithuania inhabitant). One explanation for this crimes has some grounds in these find- inverse relationship between regions’ ings, at least for vandalism in Latvia. population density and assault could be Although the variable hospital beds per that degree of urbanity affects acquisitive inhabitant does not help in explaining crimes more than expressive ones. At the variation of homicide ratios in the regional level, it is acceptable to assume Baltic countries, there are reasons to that urban areas are often more crimino- believe that improvement in health care genic than rural areas, not because they and emergency services in the 1990s concentrate lots of people per area but might be related to the reduction of because urban areas offer more opportu- deaths caused by non-natural causes in nities for crime (e.g. stock of goods) than the region as a whole. rural ones. High divorce rates also tend to explain Final considerations high vandalism ratios. As in the case of Although Estonia, Latvia, and Lithuania lethal violence, individuals from broken became independent republics in the families are more likely to be involved in early 1990s and have since experienced acts of property damage. It may be a form profound political, economic and social of entertainment among groups of young changes (including EU membership), people (Mahoney and Stattin 2000) or a very little is known about these coun- way to express dissatisfaction triggered tries’ crime levels and geography. This by ‘relative deprivation’ (Burton et al. article aims to address this gap by 1994). As Sampson and Raudensbush improving the knowledge base regarding (1999:609) claim, predatory crimes, such the levels and spatial distribution of as vandalism, may not directly ‘cause’ expressive crimes in Estonia, Latvia and other more serious crimes, but they do Lithuania between 1993 and 2000. It also share the same explanatory processes (as, provides an analysis of intra-regional in this case, homicides), with the differ- variation of expressive crime ratios in ence that they can be observed by every- relation to demographic, socio-economic body in the area: residents, visitors and and social cohesion variables after con- potential offenders. trolling for other covariates. In the case of Latvia, regions with low Estonia, Latvia and Lithuania had numbers of hospital beds per inhabitant significantly higher homicide rates than were also associated with high vandal- other European countries, but lethal ism ratios. Hospital beds per inhabitant violence is the only expressive crime was used in this analysis as an indicator analysed here that has been in decline in of how well regions cope with the the 1990s. Causes of relatively high challenge of adapting to market-oriented homicide rates in the Baltic countries principles and, at the same time, keeping include differences in their demographic some of the core welfare services, such composition (e.g. higher propensity as public health. The previously stated among certain ethnic minorities, at least hypothesis that regions with low num- in Estonia), increasing anomic condi- bers of hospital beds per inhabitant tions (e.g. following an increase in would have high ratios of expressive economic disparities), life-style (e.g.

20 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania alcohol consumption) and conflicts consequence, crime levels. In order to between organized crime groups. Whilst test how these factors affect regional police official statistics on expressive crime patterns in the Baltic states, crimes such as assault and vandalism regression models were used to explore have increased several-fold in the Baltic the significance of a set of variables as countries, victimization crime surveys predictors of expressive crime ratios in indicate that there have been no signi- the year 2000 in different regions in ficant changes in crime levels and Estonia, Latvia and Lithuania. Results composition in the 1990s. A possible show that covariates related to regions’ explanation for the mismatching between social structure (e.g. divorce rate, non- police statistics and victimization surveys native population, young male popula- is a rise in reporting to the police as a tion) are more powerful to describe consequence of increasing trust in public expressive crime ratios in the year 2000 institutions during the 1990s. An alter- than other regional (economic covariates, native reason behind the overall stability such as foreign direct investment per in victimization is the increasing polari- capita) or locational indicators (being a zation between different groups of border/non-border region). With very few society. Although data are not available exceptions, covariates function in ways to check the causes behind this stability in which would be predicted by Western the Baltic countries, there is a chance literature on crime geography. For that, as in the UK and Sweden, this is instance, divorce rate is significant in all caused by the fact that the proportion of models for expressive crimes. In Latvia, those victimized has become significantly however, divorce rate is negatively related greater among the poor and has stabilized to homicide rates. Although the mechan- among the richer groups of society. isms relating lethal violence and low Expressive crimes are often more divorce rate are unknown at an individual wide-spread in urban than in rural areas level, it is likely that at an aggregate level in the Baltic countries. Internal regional divorce rate mimics the behaviour of differences in rates were influenced by missing covariates in the model. To social disorganization risk factors, such approach the question of the mechanisms as differences in a region’s economic underlying such a relationship, it would structure and employment, social exclu- be beneficial in future analysis to integrate sion, and by life-style factors, such as individual level data, such as data on alcohol and drug abuse—factors that are offence, victim and offender. Another often more tangible in urban areas. alternative would be to investigate spe- Differences in land use, such as the cific cases of murders in Latvia to check presence of shopping centres, large whether and how they differ in nature retailers and recreational areas, as well from the ones in Estonia and Lithuania. as differences in regions’ functionality An important area of study that has (e.g. capital cities, holiday resorts, not been covered by this article is the industrial towns) and locational factors effect of organized crime on expressive (e.g. border regions), are also expected crimes in the Baltic countries. Evidence to affect human interactions and, as a shows that organized crime contributed

Journal of Scandinavian Studies in Criminology and Crime Prevention 21 ceccato: expressive crimes in estonia, latvia and lithuania considerably to the high number of alcohol consumption, would certainly homicides and other violent crimes in shed light on the nature of expressive the Baltic countries in the early 1990s crimes in the Baltic countries, particularly but that its influence waned by the end of homicides. Moreover, new modelling of the decade (e.g. Saar et al. 2003). It approaches could be tested, for instance would be useful to investigate whether by excluding the most extreme areas from there are still any links between local/ the rest of the data set to check how these regional organized crime and levels of areas impact on the results. Although this violence. article recognizes that the quality and By incorporating the spatial dimen- availability of data have improved since sion using GIS, this analysis allows an these countries’ independence (including assessment of how social contexts and both crime and socio-economic statistics) land use structures interact at a regional much still needs to be done to meet the level to produce different patterns of requirements of a rigorous long-term crime. However, one limitation with this research design. study is that the regression analysis is based on a cross-sectional framework Acknowledgements (one-year database) which might be considered too narrow a time period The support of The Bank of Sweden for drawing final conclusions on these Tercentenary Foundation (Riksbankens countries’ criminogenic conditions for Jubileumsfond, grant J2004–0142:1) is expressive offences. Because there is gratefully acknowledged by the authors. evidence that change is still occurring Thanks to colleagues Andri Ahven, Aira in many post-socialist countries, future Veelmaa (Estonia), Alfredas Kiskis, studies should consider using a long- Danguole Bikmanaite, Nijole Lukyte itudinal design, even though this type of (Lithuania) and Andris Kairiss (Latvia) research framework would be highly for providing data and support along the dependent on the availability of good project. A special thanks goes to the and comparable quality data. The incor- experts in Estonia, Latvia and Lithuania poration of new variables (not available who kindly answered the questionnaire for this analysis) in the models, such as on crime data availability and quality.

References Ahven A (2004). Public Safety and Well-being. Aasland A, Flotten T (2001). Ethnicity and Social Trends 3, Tallinn: Statistical Office Social Exclusion in Estonia and Latvia. of Estonia. Europe-Asia Studies 53:1023–49. Anselin L (2003). GeoDa 0.9.3 User’s Guide. Ahven A (2002). Crime in the Baltic States: Spatial Analysis Laboratory, Department of comparing official statistics with survey Agricultural and Consumer Economics, data. In: Nieuwbeerta P (ed). Crime University of Illinois, Urbana-Champaign, Victimization in Comparative Perspective. Urbana, IL 6180. Available at: https:// Results from the International Crime geoda.uiuc.edu/pdf/geoda093.pdf. Victims Survey, 1998–2000. p. 67–79. Den Aromaa K (1998). The International Haag: Boom Juridische uitgevers. Victimization Survey in Finland, Estonia,

22 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

Latvia, and Lithuania 1995–1997. In: Eurogeographics (2005). SABE Aromaa K (ed). The Baltic Region. 2004—Seamless Administrative Scandinavian Studies in Criminology. Oslo: Boundaries of Europe User Guide. Version Pax Forlag. 1.0, 78. Available at: http://www.ign.fr/ Berglund S, Aarebrot F, Vogt H, telechargement/P.I/ActusAgenda/Europe/ Karasimeonov G (2001). Challenges to ug2001vii_final.pdf (accessed 12 September Democracy. Cheltenham: Edward Elgar. 2007). Bernhagen P, Marsh M (2007). Voting and European Sourcebook of Crime and Criminal Protesting: Explaining Citizen Participation Justice Statistics (1996). Draft Model. in Old and New European Democracies. European Committee on Crime Problems Democratisation 14:44–72. (chair: M. Killias). Strasbourg: Council of Burton VS Jr, Cullen FT, Evans TD, Europe. Available at: http://www. Dunaway RG (1994). Reconsidering Strain europeansourcebook.org (accessed 16 Theory: Operationalization, Rival Theories March 2006). and Adult Criminality. Journal of European Sourcebook of Crime and Criminal Quantitative Criminology 35:213–39. Justice Statistics (2003). Second Edition. Butler WE (1992). Crime in the Soviet Union: European Council, Committee of Experts Early Glimpses of the True Story. British (chair: M. Killias). The Hague/London/ Journal of Criminology 32:144–66. Paris: WODC/Home Office/ESC-e˙cole des Ceccato V (2006). Crime Dynamics at sciences criminelles, 2003. Onderzoek en Lithuanian Borders. European Journal of beleid 212. Available at: http://www. Criminology 4:131–60. minjust.nl:8080/b_organ/wodc/reports/ Ceccato V, Haining R (2004). Crime in Border ob212i.htm (accessed 16 March 2006). Regions: The Scandinavian Case of Europol (2004). 2004 European Union O¨ resund, 1998–2001. Annals of the Organised Crime Report. Luxembourg: Association of American Geographers Europol 26. 94:807–26. Eurostat (2003). New Chronos Database: Chamlin MB, Cochran JK (1995). Assessing Number of Divorces per 100 New Marriages in Messner and Rosenfed’s Institutional the Corresponding Year. Available at: http:// Anomie Theory: a Partial Test. www.eurofound.eu.int/areas/qualityoflife/ Criminology 33:411–29. index.htm (accessed 7 July 2006). Christie N, Andenaes J, Skirkbeckk S (1965). A GavrilovaNS, Semyonova V, Evdokushkina GN, Study of Self-reported Crime. Scandinavian Gavrilov LA (2000). The Response of Violent Studies in Criminology 1. Mortality to Economic Crisis in Russia. Corcoran ME, Chaudry A (1997). The Population Research and Policy Review Dynamics of Childhood Poverty. The 19:397–419. Future of Children 7:40–54. Gruszczynska B, Gruszczynski M (2004). Del Frate AA, Van Kesteren J (2004). Criminal Crime in Enlarged Europe: Comparison of Victimisation in Urban Europe. UNICRI. Crime Rates and Victimization Risks. Available at: http://www.unicri.it/icvs Transition Studies Review 12:337–45. (accessed 18 November 2006). Haining R (2003). Spatial Data Analysis: Durkheim E (1897). Suicide: a Study in Theory and Practice. Cambridge: Sociology. New York: Free Press. Cambridge University Press.

Journal of Scandinavian Studies in Criminology and Crime Prevention 23 ceccato: expressive crimes in estonia, latvia and lithuania

Heiskanen M, Sire´n R, Aromaa K (2004). Messner SF, Rosenfeld R (1994). Crime and the Victimisation and Fear in Finland 2003. American Dream. Belmont: Wadsworth. Interim Report of the 2003 National Survey Messner SF, Rosenfeld R (1997). Political of Victimisation to Accidents, Violence, Restraint of the Market and Levels of Property Crime, and Fear. Research Criminal Homicide: a Cross National Communications 59. Application of Institutional Anomie Interpol (2003). International crime statistics. Theory. Social Forces 75:1393–416. Available at: http://www.interpol.int/ Nilsson A, Estrada F (2006). The Inequality of Public/Statistics/ICS/downloadList.asp. Victimisation: Trends in Exposure to Crime Kim SW, Pridemore WA (2005). Social Smong Rich and Poor. European Journal of Change, Institutional Anomie and Serious Criminology 3:387–412. Property Crime in Transitional Russia. Nordic Criminal Statistics (2003). OES British Journal of Criminology 45:81–97. Project 2001. One Europe or Several? Lehti M (1998). Recent crime trends in Eastern Project Outline: Crime, Borders and Law Europe. In: Aromaa K (ed). The Baltic Enforcement—a European Dialogue for Region. Scandinavian Studies in Improving Security. Available at: http:// Criminology. Oslo: Pax Forlag. www.one-europe.ac.uk/cgi-bin/esrc/world/ Lehti M (2001). Homicide trends in Estonia, db.cgi/proj.htm?id511. 1971–1996. In: Ylikangas H, Karonen P, Orru`M(1987). Anomie: History and Lehti M (eds). Five Centuries of Violence in Meanings. Boston: Allen & Unwin. Finland and the Baltic area. p. 133–92. Pampel F, Gartner R (1995). Age Structure, Columbus: Ohio State University Press. Socio-political Institutions, and National Leon DA, Shkolnikov VM (1998). Social Stress Homicide rates. European Sociological and the Russian Mortality Crisis. Journal of Review 11:243–60. the American Medical Association 279:790–1. Perry B (2003). Hate and Bias Crime: A Levitt S (1999). The Changing Relationship Reader. New York: Routledge. between Income and Crime Victimisation. Philipov D, Dorbritz J (2003). Demographic Economic Policy Review 5:87–98. Consequences of Economic Transition in Lewis P (1997). Political participation in post Countries of Central and Eastern Europe. communist Democracies. In: Potter D, et al. Population studies 39, Strasbourg: (eds). Democratization. p. 459. Cambridge: Directorate General III—Social Cohesion, Polity Press. Council of Europe. p. 208. Lotspeich R (1995). Crime in the Transition Pridemore WA (2000). Determinants of Economies. Europe-Asia Studies Regional Homicide Mortality Patterns in 47:555–89. Post-Soviet Russia. Paper presented to Mahoney JL, Stattin H (2000). Leisure Time IUSSP Committee on Emerging Health Activities and Adolescent Anti-social Threats, Seminar 1: Determinants of Behavior: the Role of Structure and Social Diverging Trends in Mortality. Available Context. Journal of Adolescence 23:113–27. at: http://www.demogr.mpg.de/Papers/ Mays JB (1970). Crime and Its Treatment. workshops/020619_paper32.pdf (accessed London: Longman. 11 November 2006). Merton RK (1967). Social Theory and Social Pridemore WA, Spivak AL (2003). Patterns Structure. New York: Free Press. of Suicide Mortality in Russia. Suicide

24 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

and Life-Threatening Behaviour 33: Van Duyne PC, Ruggiero V, Scheinost M, 132–50. Valkenburg W (eds) (2001). Cross-border Putnam R (1995). Tuning In, Tuning Out: The Crime in a Changing Europe. Huntigton: Strange Disappearance of Social Capital in Nova. America. Political Science and Politics Vetik R (1993). Ethnic Conflict and 28:664–83. Accommodation in Post-Communist Rawlinson P (2001). Russian Organised Crime Estonia. Journal of Peace Research and the Baltic States: Assessing the threat. 30:271–80. Working Paper 38/01. Available at: Western JS, Lynch M, Ogilvie E (eds) (2003). http://www.one-europe.ac.uk/pdf/ Understanding Youth Crime: an Australian w38rawlinson.pdf (accessed 12 September Study. Aldershot: Ashgate. 2007). Wikstro¨m PO (1991). Urban Crime, Criminals and Victims: the Swedish Experience in an Saar J, Markina A, Ahven A, Annist A, Ginter J Anglo-American Comparative Perspective. (2003). 1991–2001. Stockholm: Springer-Verlag. 2nd ed. Tallinn: Juura. (In Estonian and World Bank (2005a). The Baltic Growth English). Acceleration—is it Sustainable? Quarterly Sampson RJ (1986). Crime in Cities: The Economic Report, Vol. 3, Part III. Available Effects of Formal and Informal Social at: http://www-wds.worldbank.org/servlet/ Control. Crime and Justice 8:271–311. WDS_IBank_Servlet?pcont5details& Sampson RJ, Raundensbush SW, Earls F eid5000160016_20050615135856 (accessed (1997). Neighborhoods and Violent Crime: 17 March 2006). a Multilevel Study of Collective Efficacy. World Bank (2005b). Enhancing Job Science 277:918–24. Opportunities in Eastern Europe and the Sampson RJ, Raundensbush SW (1999). Former Soviet Union. Available at: http:// Systematic Social Observation of Public siteresources.worldbank.org/INTECA/ Spaces: a New Look at Disorder in Urban Resources/laborstudy05-fullreport.pdf Neighbourhoods. American Journal of (accessed 5 June 2006). Sociology 105:603–51. Young J, Mathews R (2003). New Labour, Savolainen J (2000). Inequality, Welfare State crime control and social exclusion. In: and Homicide: Further Support for the Mathews R, Young J (eds). The New Institutional Anomie Theory. Criminology Politics of Crime and Punishment. Devon: 38:1021–42. Willan Publishing. Stickley A, Ma¨kinen IH (2005). Homicide in the Russian Empire and Soviet Union: vaˆ nia ceccato Continuity or Change? British Journal of Department of Urban Planning and Environ- Criminology 45:647–70. ment Trickett A, Ellingworth D, Hope T, Pease K Royal Institute of Technology (1995). Crime victimisation in the eighties. Drottning Kristinas va¨g 30 British Journal of Criminology 35:343–59. S-100 44 Stockholm Ulrich C (1994). The growth of SWEDEN and the Baltic region: RFE/RL Research Tel: +46 (0)8 7908709 Report, 3, 23. Email: [email protected]

Journal of Scandinavian Studies in Criminology and Crime Prevention 25 ceccato: expressive crimes in estonia, latvia and lithuania

Appendix 1 Crime rates per 10,000 inhabitants, selected European countries (definitions of offences vary between countries, both due to legal differences and statistical recording methods; for a comparative definition, see Appendix 2)

Total recorded crime per 10,000 inhabitants:

1993 1994 1995 1996 1997 1998 1999 2000 Estonia 245 238 267 241 291 328 374 421 Latvia 155 149 156 153 149 149 181 209 Lithuania 162 157 164 183 211 219 218 235 Czech Republic 386 360 364 382 392 414 415 381 Poland 222 235 253 232 257 278 290 328 Slovenia 222 219 188 185 194 288 321 351 Norway 567 576 656 667 701 720 711 737 Sweden 1,366 1,267 1,298 1,329 1,346 1,332 1,346 1,369 Denmark 1,054 1,051 1,030 1,004 1,005 941 929 945 Finland 756 753 747 735 728 745 722 747 England and Wales 1,085 1,008 1,004 961 884 991 1,025 996 Germany 832 803 816 812 803 787 768 762 France 675 680 634 613 597 607 610 642 Italy 396 380 396 422 425 422 412 382 Netherlands 768 784 729 693 708 702 729 739 Portugal 308 330 326 320 322 341 361 362 Source: Crime data: Board, Latvia’s Ministry of Interior, Lithuania’s Ministry of Interior, Home Office (2003); population data: UNECE (2002).

26 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

Homicides per 10,000 inhabitants:

1993 1994 1995 1996 1997 1998 1999 2000 Estonia 1.77 1.99 1.65 1.43 1.26 1.39 1.07 0.97 Latvia 1.66 1.47 1.12 1.03 1.05 0.97 0.88 0.89 Lithuania 1.29 1.41 1.35 1.09 1.09 1.00 0.97 1.13 Czech Republic 0.27 0.28 0.27 0.26 0.28 0.30 0.26 0.27 Poland 0.33 0.24 0.22 0.23 0.21 0.20 0.19 0.22 Slovenia 0.37 0.17 0.23 0.19 0.19 0.08 0.13 0.15 Norway 0.23 0.08 0.10 0.10 0.09 0.09 0.08 0.11 Sweden 0.12 0.10 0.10 0.12 0.01 0.11 0.12 0.10 Denmark 0.14 0.14 0.11 0.13 0.17 0.09 0.10 0.11 Finland 0.33 0.33 0.34 0.37 0.27 0.22 0.28 0.28 England and Wales 0.26 0.25 0.27 0.26 0.27 0.27 0.29 0.30 Germany 0.18 0.17 0.17 0.15 0.14 0.12 0.12 0.12 France 0.30 0.24 0.23 0.20 0.16 0.16 0.16 0.18 Italy 0.20 0.18 0.18 0.17 0.16 0.16 0.15 0.14 Netherlands 0.19 0.18 0.19 0.15 0.15 0.13 0.15 0.14 Portugal 0.27 0.17 0.12 0.12 0.13 0.15 0.13 0.12 Source: Crime data: Estonian Police Board, Latvia’s Ministry of Interior, Lithuania’s Ministry of Interior, Home Office (2003), European Sourcebook of Crime and Criminal Justice Statistics (1996, 2003); population data: UNECE (2002).

Journal of Scandinavian Studies in Criminology and Crime Prevention 27 ceccato: expressive crimes in estonia, latvia and lithuania

Assault per 10,000 inhabitants:

1993 1994 1995 1996 1997 1998 1999 2000 Estonia 2.47 2.74 2.74 2.70 3.09 2.91 2.73 2.97 Latvia 0.50 2.81 2.39 1.89 2.05 1.75 1.77 1.76 Lithuania 2.78 3.45 3.27 3.31 3.47 4.17 4.56 4.80 Czech Republic 7.70 7.10 7.75 7.55 7.43 7.72 7.19 7.00 Poland 4.30 4.70 7.54 7.91 8.55 8.33 7.82 8.39 Slovenia 2.40 2.10 2.47 2.45 2.20 2.03 2.11 2.24 Norway n.a. n.a. 5.76 5.94 5.97 5.97 6.69 7.82 Sweden 58.4 61.1 61.62 60.65 62.17 64.14 67.55 66.32 Denmark 18.0 19.0 16.48 16.32 16.53 15.96 16.87 18.30 Finland 36.80 39.0 43.40 47.86 48.32 49.78 50.77 53.78 England and Wales 38.40 40.10 39.22 43.92 46.76 73.25 83.62 85.61 Germany 33.50 34.10 37.00 38.85 40.61 40.76 44.91 46.42 France 9.80 10.90 12.22 12.93 14.02 14.83 16.22 18.05 Italy 3.70 3.70 3.75 4.14 4.39 4.65 5.20 n.a. Netherlands 16.10 18.10 18.05 19.31 24.08 24.20 26.77 27.75 Portugal 31.60 32.50 34.83 35.52 37.18 41.04 40.84 43.32 Source: Crime data: Estonian Police Board, Latvia’s Ministry of Interior, Lithuania’s Ministry of Interior, Home Office (2003), European Sourcebook of Crime and Criminal Justice Statistics (1996, 2003); population data: UNECE (2002).

28 Journal of Scandinavian Studies in Criminology and Crime Prevention ceccato: expressive crimes in estonia, latvia and lithuania

Appendix 2 Characteristics of the data set Statistics Estonia Central Statistical Bureau of Latvia Statistics Lithuania YearLithuania: 2000 Estonia: 2000; Latvia: 2000; Lithuania: 2001 Estonia: 2000; Latvia: 2000; Lithuania: 2001 Estonia: 2000; Latvia: 2000; Lithuania: 2001 Estonia: 2000 (Nomenclature of Territorial Units for StatisticsLatvia: 3); Source 2000; Lithuania: 2000 Estonia: 2000; Latvia: 2001; Lithuania: 2000 Estonia: 2000; Latvia: 2000; Lithuania: 2001 Estonia: 2000; Latvia: 2000; Lithuania: 2000 c a b 15–29 (YoungMalePop) (NonNativePop) inhabitants (DivorceRate) per 1,000 live births inhabitants (HospitalBeds) (GDPcap) per capita in Euros Male population aged Non-native population Assault Vandalism Description Type Homicide Unemployment rateDivorce rate per 1000 Deaths under 1 year old Hospital beds per 1,000 GDP per Estonia: capita 2000, in Latvia: Euros 2000, Foreign direct investment Proportions of: Appendix 2 Characteristics of the data set Type of data Offences Demographic, socio-economic, social cohesion and welfare covariates

Journal of Scandinavian Studies in Criminology and Crime Prevention 29 ect:epesv rmsi soi,lti n lithuania and latvia estonia, in crimes expressive ceccato:

30 Appendix 2 Continued.

Type of data Description Type Homicidea Assaultb Offences Vandalismc Year Source Natural increase (NaturalInc) Estonia: 2000; Latvia: 2000; Lithuania: 2000 Net migration Estonia: 1999; Latvia: 2000; Lithuania: 2001 Voter turn-out in parliament Estonia: 1999; Latvia: 2002; Lithuania: elections (VoterTurnout) 2004 Other covariates Dummy for border regions (Border) ora fSadnva tde nCiiooyadCiePrevention Crime and Criminology in Studies Scandinavian of Journal Population density (PopDens) Estonia: 2000; Latvia: 2000; Lithuania: 2005 Geographical units 107 units for the Baltic region (Estonia: 15, 2004 Eurogeographics excluding Hiiumaa, Saare and other small islands; Latvia: 33; and Lithuania: 59. aIntentional homicide means intentional killing of a person. Where possible, the figures include: assault leading to death, euthanasia, infanticide. According to European Sourcebook of Crime and Criminal Justice Statistics (2003), Estonia and Latvia exclude assault leading to death and cases of euthanasia. bAssault means inflicting bodily injury on another person with intent. Where possible, the figures exclude: assault leading to death, threats, acts just causing pain, slapping/punching, sexual assault. According to European Sourcebook of Crime and Criminal Justice Statistics (2003), Lithuania includes assault leading to death into homicide statistics. Threats are included in Latvia, whilst acts causing pain are included in Estonia and Latvia. cVandalism is a criminal offence involving damage to or defacing of property belonging to another person or the public. It can also be defined as a ‘wilfulor malicious destruction, injury, disfigurement, or defacement of any public or private property, real or personal, without the consent of the owner or persons having custody or control’. This should include damage to public and private property.