„European Journal of Criminal Policy and Research”, vol. 15 iss. 4 (2009), s. 327.

I.     Introduction

This paper applies the model created within the Mainstreaming Methodology for the Estimation of the Costs of Crime (MMECC) project.[1] The project has aimed at creating a model that can be applied to estimate the cost of crime in Europe and elsewhere, basing on the fast growing research on the cost of crime.[2] In this paper, using crime and other data for Poland, I explore the feasibility of building such a model. This includes identifying the data requirements and discussing alternative data sources. The paper provides the first estimates of the cost of crime in Poland.[3] In estimating the costs of crime I follow the bottom-top approach, which gathers different elements of costs of various crimes and put them together and is prevalent in the literature.[4]    Measuring crime is itself a difficult task. Official police statistics cover only part of all crimes committed, for victims are not necessarily likely to report the fact of victimisation to the Police, and the Police do not register all reported incidents. The shortcomings of official statistics have prompted the development of different means to measure crime, in particular the use of victimization surveys, i.e. public surveys that ask samples of the population about whether they have been victim of particular kinds of crime. In Europe, the International Crime Victim Survey (ICVS) was launched in 1988 to measure the extent of victimization for 11 crimes. Poland has taken part in this project since 1991. Victimisation data for 2003 from this source are used in the paper[5]. The ICVS dataset covers assaults and threats, robbery, rape, sexual assaults, car theft, theft from the car, bike theft, personal theft and burglary in a dwelling. Burglary of a second home was added to the Polish version of the questionnaire. The survey does not, obviously, cover homicides and other crimes resulting in deaths. For including crimes resulting in death, this paper uses the police statistics. As the survey covers only the population aged 16 years and over, the rates reported relate only to that population, leaving aside substantial youth victimisation.[6] The ICVS does not cover crimes against businesses. This is a non-trivial part of all crimes, in particular those against property. Police statistics often report crimes in terms of the number of incidents of crime, not the number of victims. However for calculating the costs of crime the latter is more important, and usually higher, for some crimes have multiple victims. But in some cases involving multiple offenders each offender’s behaviour may be recorded as a separate crime.  This is similar to the conviction statistics, where the number of convictions is based on the number of offenders, not crimes nor victims. Only completed crimes are counted here, and attempts are excluded. Unsuccessful attempts obviously lead to lower costs than completed crimes, so it would be inappropriate to treat them on the same footing, even if they are often included in the total number of crimes/convictions. Values are expressed in 2003 PLN (Polish Zloties.) The average yearly exchange rate for 2003 was 1 Euro = 3.8898 PLN. 

II.   Violent crimes

Among crimes against individuals, violent crimes are intuitively the most important. The ICVS covers a broad range of violent crimes. The broad category of violent crimes has been divided into three categories: loss of life (based on police statistics), assaults and robberies, and sexual crimes (data taken from the ICVS).

1.   Loss of life

This category is not based on the ICVS data for obvious reasons. While homicide is the most obvious example of a crime resulting in death, there are other crimes with lethal results. In the Polish Criminal Code, there are 27 types of crimes with lethal results: apart from 6 different kinds of homicide (for example, homicide committed with special cruelty, homicide upon provocation, infanticide, mercy killing), there are assaults leading to death (death as a negligent result of an offender’s intentional assault), negligent manslaughter (for example resulting from not observing the health and safety regulations at work place), and – the most numerous – cases of dangerous driving that resulted in car accidents with casualties. The full list of legal classifications and the number of crimes and victims in 2003 is provided in Table 1.   [Table 1 about here] As clearly seen in Table 1, criminal loss of life may be caused by very different behaviour. While homicide is a clear cut example of a criminal loss of life, there are other crimes resulting in death as well. Some of them are commonly criminalised in many countries, like death by dangerous driving or assaults leading to death. Others, like abortion or euthanasia, may be legal in some jurisdictions. Therefore, in analysing and comparing costs of crime among countries, one may choose either to include only those crimes that are common for all, or most, jurisdictions, or to follow the particular definitions of criminal behavior used in each country. The latter approach has been used here. The reason is that estimates of the costs of crime in any given society should include (and be limited to) the costs of behaviour defined as criminal by that society. This makes the assumption that criminal legislation reflects appropriately the preferences of a given society into account. It may not always be exactly true but is considerably preferable to any alternative.  Although death may be a common result of many crimes, punishment for them varies considerably, reflecting different degrees of culpability on the part of offenders. One may question then whether it is appropriate to assign the same amount of harm for these categories: for example, while death is the result equally of homicide and of causing death by dangerous driving, the amount of sorrow and regret of the victim’s relatives may be higher in the former than in the latter case. The amount of harm is likely to be context-sensitive – it is relevant whether someone has his arm broken because of an accident, or because of a deliberate attack. As estimates in the article are based on values from health economics, they are likely, because they have been derived in a more morally neutral setting,  to underestimate the amount of harm flowing from the criminal deeds.The valuation of human life originated 40 years ago,[7] but remains a controversial issue in the literature. It is true that human life is priceless, and it would offend our moral sense if someone attached a price tag to it. From the other side, each of us, and society in general, makes decisions every day that influence the probability of surviving. The amount of resources devoted to health care, road safety, or product safety has a direct impact on the number of deaths. So has our pattern of behaviour: risky jobs, extreme sports, going out at night, diet habits, all these have some influence on our life expectation. Political decisions about resource allocation and the level of regulatory activity in risky industries usually implicitly assume some value of human life which justifies the costs of implementing the prevention measures. Assigning the value of statistical life in an explicit way facilitates comparisons of different programs and allows shifting of resources in a way that saves the highest number of lives at any given budget. The same reasoning applies as well to injuries.In order to assign a monetary value to criminal deaths and physical injuries, a QALY (quality adjusted life years) approach has been used here. The QALY methodology translates every physical injury into losses in the quality and quantity of further life. For example, if after suffering an injury a victim spends one year in half health, this translates into the loss of 0.5 QALY. Death is assigned the value of 0: in such cases, the QALY loss equals the number of years a victim would otherwise survive. It is assumed here that the victims of crime are in full health, so the quality of their prospective years would be 1, and therefore the QALY loss simply equals the number of years a victim could have expected to survive.    The mean (and median) age in Poland was 36 years in 2003. Life expectancy at this age was 40 on average.[8] Therefore, it is assumed that each loss of life constitutes a loss of 40 years of life. This value is based on mean values for society, while in the case of crime, the distribution of victims’ age may be different. There are two reasons why mean values have been used here. The first (and sufficient) reason is that there are no data on victims’ age in such cases in Poland. The second reason is more normative: it is questionable whether one is allowed to (or would want to) assign different amount of harms according to victims’ age, sex, occupation, social status, and so on. It would lead to different values for children and elders, men and women. Such an approach would violate the legal principle of equality, deeply embedded in constitutional and criminal law. Therefore, it is assumed here that the society chooses to treat each loss of life equally and is willing to pay the same amount of money to prevent it.[9]While for many purposes the severity of health injuries may be comparable on a QALY basis alone, a monetary value of QALY has to be found for making any cost-benefit analyses. There are two possible sources of such a monetary value of one QALY: first are willingness to pay surveys, in which people are asked about the amount of money they are willing to spend on some health improvement (which has a known QALY equivalent.) In the only study on the willingness to pay conducted to date in Poland, Ami et al (2006) report that the implicit value of a QALY is between 40,000 and 80,000 PLN (in the context of an air pollution reduction programme which would lower the mortality rate.)The second possible source of the monetary value of a QALY is the normative decision of relevant authorities, for example in the practice of national health systems. The Polish Agency for Health Technology Assessment has adopted the WHO recommendations that the value of one QALY should be in the range 1 to 3 times GDP per capita in a given country. GDP per capita in Poland in 2003 was 22,078 PLN, and therefore the acceptable value of a QALY, according to WHO recommendations, would be 22,078 to 66,234 PLN. For purposes of this paper, a QALY value of 60,000 PLN has been adopted, which is a point estimation of the WTP survey, and remains in line with the WHO recommendations. Applying a 2% discount rate, the total present value of preventing one fatality (loss of 40 QALYs) is 1.674 mln. PLN. [Table 2 about here]            The total cost of criminal deaths was estimated at 7.6 billion PLN, with deaths resulting from dangerous driving being the most numerous and costly sub-category.

2.   Assault and robbery

In addition to criminal deaths, other violent crimes, like assaults and robberies, often have a dramatic impact on victims. Injuries with long-lasting health consequences as well as immediate trauma effects flowing from victimisation severely influence the well-being of victims. As the Police statistics do not correctly reflect the real frequency of assaults and robberies, the ICVS data are used to approximate the number of incidents. Unfortunately, the ICVS survey does not provide much information about the characteristics of assaults. In particular, it does not ask respondents about the injuries they suffered. What it does provide is an answer to the question whether there were any injuries, and whether a respondent sought medical help. The ICVS question concerns both assaults and threats. Only those cases where force was actually used are investigated here, so threats are disregarded. Altogether, in the sample, there were 149 victims of assaults where force was actually used in 2003 (2.97% of the sample). 88 of the respondents reported having been victimised once, 25 twice, 15 three times, 3 four times, and 17 reported being victimised 5 times or more. In this latter case, the number of victimisations was assumed to be 5.) Altogether, the respondents reported having been victim of 280 incidents. Injuries were incurred in 59.7% of incidents where force was used. The question about injury relates only to the last incident. Assuming the same ratio for all cases, there were 167 incidents which resulted in injuries. In 38.2% of cases where there were injuries, the victim went to see a doctor (again, the survey question related to the last incident only). Assuming the same ratio, there were 64 incidents of assults which resulted in injuries, and the victims saw a doctor. In the case of robbery, there were 63 victims in 2003 (1.26% of the sample). 51 of them reported having been robbed once, 10 twice, and 2 three times. Altogether, there were 77 incidents of robbery. Injuries were incurred in 24.7% of cases (19 cases). In these cases 44.3% victims sought medical help (8 cases). Moreover in 39.8% of cases, an offender succeeded in stealing some property from the victim. As has been previously noted, the ICVS does not provide an injury profile for violent crimes (assaults, robberies). To provide estimates of the losses, Dolan et al. (2005) estimates were used. Dolan et al. (2005) based their estimates of QALY loss for each crime category on the injury profile revealed by the British Crime Survey. The following QALY losses were reported: common assault 0.007, other wounding 0.031, serious wounding 0.191, robbery 0.028. While there is no possibility of comparing injury profiles directly, we may use the pattern of answers to the questions that are the same in both surveys: whether there were injuries, and whether the respondent sought medical help. For common assaults, 38% of victims suffered injuries, but only 2% sought medical help. For wounding, 96% suffered injuries, and 39% went to a doctor. For robbery, 37% suffered injuries, and 19% sought medical help (Simmons et. al. 2002, Table 6.07.)             It seems that common assault may be comparable with Polish assaults in that there is (almost) no injury: only 2% of English victims sought medical help.  The QALY loss is almost entirely due to the emotional impact of the crime. Assault with injuries, but no medical help, is assumed here to be equivalent to the English category of ‘other wounding’. Accordingly, those cases in which victims suffered injuries and sought medical help are treated as ‘serious wounding’. For robbery, as the percentage of victims requiring medical help is almost half that for their English counterparts (10% to 19%), it is assumed here that the QALY loss in Poland is half of that in England and Wales, i.e. 0.014. Additionally, in the case of robbery the ICVS questionnaire used in Poland has been modified in order to include the question about the value of items stolen. The average reported value of loss was 572.33 PLN (in those 39.8% of cases when an offender actually stole something.)  Table 3 summarizes calculations: [Table 3 about here] The total cost of assaults and robberies has been estimated at 6.5 billion PLN, i.e. one billion less than the cost of criminal deaths. Among assaults, those with injuries where the victims sought medical help have been the most costly, even if not so numerous: they comprise 18% of cases, but almost 70% of the total costs of victimisation.  Common assaults (physical force used, but no injuries) have a negligible impact.

3.   Sexual crimes

Sexual crimes are commonly assessed to be one of the most traumatic crime experiences for victims. Consequently, in terms of costs they rank high in the list of most costly crimes in any cost of crime estimates. However, the frequency of such crimes is hard to measure: not only victims are very reluctant to report these crimes to the Police but, even in victimisation surveys, respondents may decline to admit having been sexually assaulted. In the Polish sweep of the ICVS 32 women[10] declared themselves to have been sexually assaulted, and the total number of reported incidents was 57.[11] The question about the details of the offence concerned only the last incident: there were 2 rapes, 4 attempted rapes, 10 sexual assaults, and 16 incidents of offensive behaviour. Assuming the same ratio of repeat victimisation (1.8) for all sub-categories, it may be estimated that there were, on average, 3.6 rapes, 7.2 attempts of rape, 17.8 sexual assaults, and 28.5 incidents of offensive behaviour. As in the case of assaults the ICVS does not provide many details about the injuries or trauma suffered. Therefore, the valuation of these consequences is based on Dolan et al. (2005) estimates. Sexually offensive behaviour is excluded from the following calculation, and attempts of rape are treated as sexual assaults rather then completed rapes. According to this study, 37% victims of rape and 14% victims of sexual assaults suffered some injury (p. 973-974.)  In Poland, both victims of rape incurred injuries, as well as 3 out of 4 victims of attempted rape, and 2 out of 10 victims of sexual assaults. It seems then that applying the Dolan et al. (2005) estimates will provide rather conservative valuation of trauma associated with rape and sexual assaults in Poland. Table 4 reports the estimated number of incidents and their costs. [Table 4 about here] The total cost of sexual crimes has been estimated at 2.2 billion PLN.  While rape is obviously more costly per case, due to the much higher number of sexual assaults, the latter is the most costly sub-category. Having in mind the very sensitive nature of such crimes, and the reluctance of respondents to admit being victimised, the numbers reported here likely underestimate the true burden of sexual crimes. However, the small size of the sub-sample of sexual crime victims in the ICVS prevents firm conclusions about the real scope of sexual abuse in Poland.  

III.  Property crimes

There is some disagreement in the literature about whether the value of stolen goods is a cost (external cost approach)[12] or just a transfer that does not constitute a cost from the social point of view (social cost approach).[13] This is part of a broader discussion about who has standing in cost benefit analysis, and, in particular, whether criminal gains should be included in social utility function. Elsewhere, I argue that only those benefits that one is willing to pay for should be included, as the consent (at least hypothetical) of the other party is a neccessary condition for the exchange to be utility-maximizing.[14] This argues against including the sadistic pleasures of offenders stemming from crimes, as well as treating the value of stolen goods as a mere transfer, not as a cost. The original ICVS questionnaire did not provide data about the losses due to crime apart from losses due to burglary. However, the Polish questionnaire for the 2003 sweep was modified so as to include questions about the value of stolen items in all property crimes. There were 8 such crimes: car theft, theft from a car, bike theft, motorbike/moped/scooter theft, burglary, robbery, personal theft, and second home (cottage) burglary. As there were just 4 cases of motorbike/moped/scooter theft in the sample, this category has been excluded from the calculations.The respondents were asked about the value of stolen items. In the case of car theft, only those respondents who did not recover their cars were asked. Damages were generally not covered – only in the case of burglary were respondents asked about the value of losses, including the value of damages. The value of stolen items in cases of robbery is reported here for comparative purposes. While, according to the legal classification of the Polish Criminal Code, robbery is an offence against property, from a criminological point of view this crime is rather a violent one.  Therefore, estimates of QALY losses for robbery are reported above, among other violent crimes, and the value of stolen items is included there as well. To avoid double counting, Table 5 reports the total with and without robbery.  [Table 5 about here] Among property crimes, car theft is by far the most costly crime, mainly due to the high value of the loss per case, followed by personal theft (which is also the most frequent property crime), and theft from a car. The estimates cover mainly the value of stolen items, therefore providing only lower bound estimates of the total losses: damage is not covered, and the costs of time lost due to the crime incident and any subsequent formalities are also excluded. It is also quite obvious that even in the case of property crimes, there are some emotional costs to victims. Such costs are not covered here. The values reported are losses of individuals only. As for crime against business, it was estimated that in Poland in 2006/2007, large retailers alone lost more than 4 billion PLN due to theft.[15] There are also substantial losses of business and public bodies due to vandalism, car theft and the like. The values reported in the Table 5 are basically the value of lost goods. Only in the case of burglary did respondents evaluate the losses and damage related to the crime together. According to the previous sweep of ICVS (in 2000), when these questions were asked separately, about 78% of the total loss was a value of lost goods, while 22% was the value of damage. With this in mind, one may calculate the value of the illegal market in stolen goods. The benefits to thieves, fences and others involved in trafficking in stolen goods amount to about 2.9 billion PLN. The value of retail sales in 2003 was about 400 billion PLN (Concise Statistical Yearbook of Poland 2004, p. 190), so – including the losses of large retailers – it seems that trafficking in stolen goods is equivalent to about 1.7% of total sales.  

IV. Costs in response to crime

Because crime is so costly, society undertakes many actions to keep the crime level low. One of these actions is the costly process of apprehending, adjudicating and punishing offenders. This requires maintaining a Police Force (although some Police tasks are preventive, i.e. they aim at preventing crimes, not at investigating them), a prosecution service, criminal courts and probation and prison services. The costs of the criminal justice system are reported below. Unfortunately, the costs of the criminal justice system in Poland are not reported according to crime category. Therefore, it is not known how much the Police, prosecution service or courts spend on investigating homicide, robbery, or theft cases. In order to assess the costs incurred by the criminal justice system in investigating the ICVS crimes (including criminal death cases), a two-stage procedure has been used.First, the share of the ICVS crimes in the total number of crimes and cases, recorded by the Police and brought before the court in 2003, has been calculated. The ICVS crime definitions were matched with the most appropriate legal definitions used by the criminal justice system.  Property crimes against businesses were excluded so far as was possible due to the limitations of data. The ICVS crimes constitute around 25% of all crimes recorded by the Police and pursued by the prosecution service in 2003. The rest of recorded crimes are mainly: some crimes against individuals not covered by the ICVS (for example vandalism, threats, not paying child support), property crimes against businesses (thefts, burglaries), business and tax crimes (fraud, falsifying documents, tax evasion), drug crimes, and traffic crimes (mostly driving while intoxicated.)Slightly less, 23%, is the share of the ICVS crimes in the total number of convictions, i.e. in the workload of courts. The share of the ICVS crimes in executing sentences is harder to calculate. There are no separate statistics on the costs of the probation service: they are counted as part of court costs. However, the costs of prisons are known. In the total burden of imprisonment[16] the ICVS crimes had a 58% share. The higher share of the ICVS crimes in prison than in the police budget indicates that the ICVS crimes are more serious than average. This is understandable as they cover violent crimes, which are punished more severely than others.Secondly, particular types of crimes have been allocated costs within the budget of a given criminal justice agency according to their relative seriousness. Seriousness has been measured by the burden of imprisonment, both suspendend and unconditional (see n. 16). This undervalues the seriousness of crimes punished mostly by community sentence or fine. However, as the most common punishment in Poland is suspended imprisonment (approx. 60% of all sentences) and unconditional imprisonment constitutes another 10%, such an approach may provide a rough estimate of the costs of executing sentences. For the ICVS crimes (which seems to be more serious than average), unconditional imprisonment was imposed in 20% of case, and suspended imprisonment in 65%. Therefore, for that group of crimes, the prison costs may be used as a proxy of costs of executing sentences in general.  It is much less clear whether allocating costs according to the seriousness of crime is a good proxy for the actual police, prosecution, and court costs. However, such values, even if not correctly reflecting the actual spending structure, may be used as a normative statement what would be the desired allocation of resources. Within a given budget it seems optimal to make allocation decisions in such a way that marginal benefits of investigating, prosecuting, and adjudicating different kinds of crimes are equalized. Otherwise, it would be possible to increase the benefits without increasing the costs. The benefits, for example the number of convictions, should be weighted by the seriousness of crime, which in turn may be measured by the actual severity of sanctions, or by the costs of these crimes to victims.[17]Table 6 estimates the costs of criminal justice system. The second column, named ‘% attr.’, shows the percentage a given crime accounts for in the total burden of imprisonment. Because it is not always possible to match the ICVS categories with legal definitions used by the criminal justice system, the burden of imprisonment represented by the broader categories of sexual crimes, theft and burglary has been used. However, for calculating cost per case, an approximation of the ICVS crimes going through the criminal justice system has been made. The estimates of the costs are based on the costs of each criminal justice agency and the number of cases going through it. As the Police record many crimes that are not pursued by the prosecution service or taken before the court,[18] and the courts do not convict all accused,[19] and not all verdicts are sentences of imprisonment, the number of cases at each stage varies. Therefore, the cost per case in court is of those cases that actually went before the court. However, the total cost per case is calculated on the basis of all crimes recorded by the Police.  That is the appropriate figure because it allows for the fact that only some proportion of crimes will proceed through the courts and make use of prison services. [Table 6 about here] The total costs of the criminal justice system devoted to ICVS crimes (including criminal loss of life) have been estimated at 3 billion PLN.  The costs of the Police and the prison service are the most important, while the costs of prosecution and adjudicating are much lower. As for particular crime categories, burglaries, robberies, and homicides are the most costly in terms of the reaction of the criminal justice system, but for different reasons. Homicides are punished by the lengthiest sentences of imprisonment so, although there are few cases, the cases impose considerable costs. On the other hand, burglaries and robberies are punished much more leniently, but there are so many instances of these crimes that the total cost is also high. Still, the calculations of the cost of criminal justice system per case are only crude approximations, and more research is required to check whether the actual costs per case are close to these estimates. However, the aggregate costs of criminal justice system are certain, so the only question is about the allocation of the budget in practice. 

V.  Summary

Table 7 summarizes the costs of the ICVS crimes, both for victims, and for the criminal justice system. For comparison, the total cost of crime is also expressed as a percentage of GDP.  [Table 7 about here] Violent crimes are the most costly. The importance people attach to their health, in terms of their willingness to pay for averting health losses, is so great that the losses due to violent crimes outweigh the property losses. Per incident, death by dangerous driving is the most costly single crime category, calling for more efficient counter measures.[20] Assaults, in particular those that led to injuries and required medical assistance, are the next most important category. In general, programs aiming at reducing violence are presumed to be the most promising in terms of their cost-effectiveness, for violence imposes a substantial cost on victims. This result is in line with other cost of crime research.Analysing the total costs of crime, it seems obvious that the costs of the criminal justice system itself, the costs borne in response to crime, constitute just a small fraction of the total. Victimisation costs are six times greater – on average – than resources spent on the police, prosecution service, courts, and prisons put together. While this fact alone does not suggest whether the current level of public spending on crime prevention is optimal or not, it does imply that in the discussion about costs of crime, public spending should not be the sole concern. It also suggests that the allocation of criminal justice resources might not be optimal. For the ICVS crimes, property crimes have a share of more than 60% in the number of convictions (which may be a proxy of resource allocation), but only a share of about 15% share in victimisation costs. The relative importance of particular crime categories is shown in the following table. [Table 8 about here] It is obvious that the criminal justice system is flooded with property crime cases, whereas their impact on victimisation costs is not so substantial. This may mean that reallocating resources toward violence-related programs, including better investigation and prosecution, may decrease the burden of crime on society. While homicide is always a priority for law enforcement agencies, so should be other violent crimes, including sexual ones. The above calculations obviously underestimate the total costs of crime in Poland. They do not cover costs borne in anticipation of crime (costs of special doors, window bars, security services, precautionary behaviour), nor costs of crime for businesses, nor costs of some crime to individuals. These other crimes are of very different types, ranging from simple shoplifting to drug offences to sophisticated fraud schemes. In 2003, the Police registered 1.47 m. crimes, of which 0.37 m. were the ICVS crimes. Excluding traffic offences (which impose risk on others, but unless there is an accident, do not cause any harm), there were 0.93 m. other crimes. If the ratio between crimes actually committed and those registered by the Police is the same for this sub-group (for the ICVS crimes the ratio is approx. 8%), and we assume the same cost per case as the cost of the ICVS property crime, the victimisation costs would almost double, adding a further 16.8 b. PLN in losses. As mentioned previously, 25% of this amount would be losses to large retailers due to theft.  [Table 9 about here] The estimated total burden of crime is 5.1% of GDP. Although a part of these estimates is highly speculative, the results are in line with the estimates for other countries. This suggests that the methodology developed within the MMECC project is able to provide meaningful results. The results are country-specific, but some values may be transferable to other countries, particularly those with similar socio-economic characteristics. For example, the victimisation costs of violent crimes may be recalculated using another country’s valuation of a QALY, and property crimes costs may be adjusted to capture the differences in GDP per capita levels. As a first approximation even those countries that have not conducted victimisation studies may use the ratio from another country of recorded crime to victimisation rates in order to assess the order of the costs of crime. The estimates developed in this article may and should be used for making cost-benefit analyses of crime policies. In the context of crime prevention, victimisation costs should be used as a measure of people’s valuation of crimes averted, in order to assess the cost-effectiveness of implementing a program. Alternatively, when there is a reduction in public expenditures (for example by implementing or extending an early release program), the likely increase in crime should be valued to check whether the savings outweigh the additional burden. While cost-benefit analyses will never be the sole criterion for decision making in the crime policy area, they nevertheless inform the policy makers about the social consequences of their actions, and should be a necessary step in any decision-making process.  AcknowledgementsI would particularly like to thank Roger Bowles and Richard Dubourg for many valuable comments they provided during their visit to Poland in November 2008  BibliographyAmi D., Bartczak A., Chilton, S., et al (2006). Final report on the monetary valuation of mortality and morbidity risks from air pollution, a part of the NEEDS project done under the 6^th Framework Programme, available at http://www.needs-project.org/docs/results/RS1b/NEEDS_RS1b_D6.7.pdf Accessed 30 January 2009.Anderson, D. A. (1999). “The Aggregate Burden of Crime”, Journal of Law and Economics 42, p. 611-642. Becker, G.S. (1968), “Crime and Punishment: An Economic Approach”, Journal of Political Economy 76 p. 169-217.Brand, S. and R. Price (2000). The economic and social costs of crime. London, Home Office.Cohen, M. A. (1988). „Pain, Suffering, and Jury Awards: A study of the Cost of Crime to Victims.” Law and Society Review 22, p. 537-555.Cohen, M. A. (2005). The Costs of Crime and Justice. New York, NY: Routledge. Cohen, M. A., Valuing Crime Control Benefits Using Stated Preference Approaches (March 2007). Vanderbilt Law and Economics Research Paper No. 08-09. Available at SSRN: http://ssrn.com/abstract=1091456The Concise Statistical Yearbook of Poland 2004 (2004), Warszawa: The Central Statistical Office.  Czabański, J. (2008). Costs of Crime Estimates and their Implications for Criminal Policy. Berlin-Heidelberg: Springer.Dolan, P., G. Loomes, et al. (2005). „Estimating the Intangible Victim Costs of Violent Crime.” British Journal of Criminology 45(6), p. 958-976. Dubourg, R., J. Hamed, et al. (2005). The economic and social costs of crime against individuals and households 2003/04. Home Office On-Line Report. 30.Gruszczyńska, B. (2008), Juvenile victimization. In M. Steketee, M. Moll, A. Kapardis (Eds.), Juvenile delinquency in six new EU member states, (pp. 103-113). Utrecht: Verwey-Jonker Instituut. Mayhew, P. (2003). Counting the Costs of Crime in Australia. Australian Institute of Criminology Technical and Background Paper Series. 4.Miller, T., M. A. Cohen, et al. (1996). Victim Costs and Consequences: A New Look. Washington, DC, National Institute of Justice.Palle, C. and T. Godefroy (2000). The Cost of Crime: A Monetary Assessment of Offending in 1996. Research on Crime and Criminal Justice in France: Penal Issues.Roper, T. and A. Thompson (2006). Estimating the Costs of Crime in New Zealand in 2003/4., New Zealand TreasurySchelling, T.C. (1968), “The life You Save May Be Your Own” [in:] Samuel B.C. Jr (ed.) Problems in Public Expenditures Analysis, Brookings Institute, Washington D.C.Siemaszko, A., Gruszczyńska, B., Marczewski, M. (2008), Atlas Przestępczości w Polsce 4, Warszawa: Oficyna Naukowa.Simmons, J. et al. (2002), Crime in England and Wales 2001/2002, London: Home Office. Van Dijk, J.J.M., van Kesteren, J.N. & Smit, P. (2008). Criminal Victimisation in International Perspective, Key findings from the 2004-2005 ICVS and EU ICS. The Hague, Boom Legal Publishers.
 Table 1 Loss of life

  
Table 2 Costs of criminal deaths (in PLN)

Notes: number of victims based on the Police statistics. 
Table 3 Costs of assaults and robberies

Notes: Number of incidents based on victimisation survey. Totals may not sum up due to rounding. The robbery’s loss includes the value of stolen goods. 
Table 4 Costs of sexual crimes

 
Table 5 Costs of property crimes

Notes: * Car theft: only those cases where a car was not recovered; ** Burglary: value of damages included (attempts of burglary excluded); *** Robbery: included only those cases where something was actually stolen.  

 Table 7 Total cost of crime

 
Table 8 Relative importance of crime categories

 
Table 9 Total cost of crime in Poland (against individuals and businesses)