Methodology - The Economic Impact of Firearm-related Crime in Canada, 2008

3. Methodology

The present study seeks to determine the consequences (category and magnitude) of firearm-related crime in Canada and estimate the associated social and economic costs.

Despite best efforts to account for all the potential consequences as a result of firearm violence, data are not available in many areas which put a significant constraint on our estimation. For example, individuals and organizations may use home security systems, guard dogs or weapons to avoid crime or for self-defence. Victims may develop certain mental health problems, such as depression, Post-Traumatic Stress Disorder (PTSD), substance abuse and suicidal behaviour. These costs are not included in this report due to the unavailability of data. Therefore, what we present here, of necessity, would not be a complete picture of the true range of costs resulting from firearm-related crime. For many cost categories that are examined in this study, assumptions have to be made. For example, because all calls made to crisis lines by victims are anonymous, no official information is available in regard to an individual's frequency of use. However, according to crisis line workers, people do make follow-up calls; therefore, we assume that on average, each victim made 5 phone calls.

3.1 Scope of the Study

In this study, the term “firearm-related crime” refers to crime where a firearm was present or used during the commission of the offence. While there are primary, secondary, and even tertiary victims, our focus will be on the primary victim of firearm-related violence. Although the impact of the violence on those who are not immediately involved will not be captured, certain important implications on family members, and other persons who were injured or threatened in the incidents will be examined. Victim costs will be gender-disaggregated to reflect the differences between male and female victims.

Two surveys - the Uniform Crime Reporting Survey 2 (UCR2) and the 2009 General Social Survey (GSS) on Victimization, cycle 23, will be used as the primary data sources to provide information on the incidents/prevalence of firearm-related violence in Canada. While the UCR2 is police-reported data recording the number of incidents that have come to the attention of police, the GSS is self-reported recording personal experiences of criminal victimization, regardless of whether or not the incident was reported to police.

Offence categories for firearm-related violence under the UCR2 are defined by the Canadian Criminal Code, including, but not limited to, homicide, attempted murder, sexual assault, assault, robbery, firearms use of, discharge, pointing, criminal harassment, uttering threats and other violent violations. On the other hand, the GSS only measures four types of violent crime that can be defined by the Criminal Code. These offences are sexual assault, robbery, attempted robbery and assault.

3.2 Categories of Costs

Drawing on a variety of methods documented in previous studies and an examination of various government and non-governmental reports, a comprehensive costing model is developed that attempts to outline the full extent of the financial costs associated with firearm-related violent crime. The costs examined herein will be broken down into three major categories Criminal Justice System Costs, Victims Costs and Third-Party Costs. Under each of these categories, we have also delineated several sub-categories to capture the economic implications of the violence. A full illustration of detailed criminal justice system cost categories are presented in Table 2.

Criminal Justice System Costs

Table 2: Categories of Criminal Justice System Costs

  1. Police
  2. Court
  3. Prosecution
  4. Legal Aid
  5. Corrections
    • 5.1 Incarceration
    • 5.2 Conditional Sentence
    • 5.3 Probation
    • 5.4 Fine

The criminal justice system plays the role of deterring, attending to, and punishing crime. There are multiple stages in the Canadian criminal justice system to deal with crimes that have been reported to police. They include police, court, prosecution, legal aid and corrections. Only a few previous studies attempt to estimate these costs system-wide (Walby 2004; Brand and Price 2000). Following Cohen, Miller and Rossman (1994), we estimate the cost of each stage separately, as the number of cases/persons involved in one stage is different from the others.

In estimating police costs, the UCR2 survey is used as the main data source as it accurately records the number of incidents of firearm-related crime that have come to the attention of police. Although police-reported data does not capture the actual number of offences (allegedly) committed in Canada, these numbers can indicate the actual resources that have been allocated to address the problem of crime. The money spent by police is mainly based on the number of incidents that have come to the attention of police, not the number of offences committed. On the other hand, the Adult Criminal Court Survey (ACS) and the Youth Court Survey (YCS) will be used to obtain the information regarding the firearm-related cases processed in the court system and court outcomes. We also use court caseloads as the base to estimate the prosecution costs and the legal aid costs.

While police and court costs appear relatively straightforward to measure as aggregate costs are available from government official sources, the cost per crime is not always available. In other words, knowing how much we spend on police or court in total tells us little regarding the appropriate allocation of dollars across different crimes, and hence, for the purposes of this study, is virtually not useful. For instance, it is difficult to determine how much of the total police expenditures in 2008 were spent on robbery with a firearm versus other robbery. And even for the same type of crime, the cost might vary considerably according to the seriousness of each incident.

The same problem also exists for estimating the costs of prosecution and legal aid. While the aggregate expenditures on criminal matters are available, there is no information to indicate the actual number of cases of firearm-related crime that were involved with prosecution and legal aid, as well as the associated cost per case if involved. In addition, offenders might hire their own defence counsel, and again it is difficult to know how much such services would cost in each case.

As a result, a realistic cost determination should be individual-incident based. For example, to accurately determine the police costs for firearm-related crime, we would require information regarding, for each single incident, the number of police officers involved, investigation time spent per police officer, their wage rates and other expenditures for items such as transportation and utilities, etc. Unfortunately, we are unaware of any such detailed data existing in Canada. Therefore, in estimating the police costs, we assign the police expenditures among different crimes according to their severity weights. In terms of court, prosecution and legal aid costs, we use the average cost per case as the estimation basis.

With respect to the cost of corrections, the estimation is relatively straightforward due to the fact that the Canadian Centre for Justice Statistics (CCJS) continuously releases the average daily cost of provincial and federal incarceration by gender, and this average cost does not vary dramatically by offence type. A breakdown of the categories associated with victim costs are provided in Table 3.

Victim Costs

Table 3: Categories of Victim Costs

  1. Health Care
    • 1.1 Physician
    • 1.2 Emergency department
    • 1.3 Acute Hospitalization
  2. Productivity Losses
    • 2.1 Lost Wages
    • 2.2 Lost Household Services
    • 2.3 Lost School Days
    • 2.4 Lost Child Care Services
    • 2.5 Lost Future Incomes
  3. Personal Costs
    • 3.1 Stolen/Damaged Property
    • 3.2 Legal Services
    • 3.3 Counselling Services
  4. Intangible Costs
    • 4.1 Pain and Suffering
    • 4.2 Loss of life

The most direct impact of crime is experienced by victims. Many costs are incurred as a direct result of firearm-related crime, such as health care costs, productivity losses and value of stolen/damaged property. While police-reported statistics are useful for understanding the nature and extent of firearm-related crimes that are reported to police and hence, the associated expenditures spent by the criminal justice system, it is noteworthy that the majority of the incidentsFootnote 11 do not come to the attention of police for various reasons, and the impact of such victimizations (not reported to police) has no less of an effect on victims and on society at large. Therefore, in order to capture a more fulsome and accurate reflection of the prevalence of firearm-related violence in Canada and hence, the impact on victims and society, self-reported victimization information from the GSS will be used.

In terms of health care costs, note that the injuries considered in the present study are not necessarily caused by a firearm. For instance, victims may suffer blunt-force trauma or stab wounds during the firearm-related incident where a firearm was not discharged. Since the GSS does not have a variable to distinguish whether the injuries were firearm injuries or non-firearm injuries, both types of injuries will be examined, unless otherwise indicated. This decision is appropriate for the purposes of this study in the sense that all the injuries are the consequences of firearm-related crime.

Nevertheless, it is also interesting to know the health care costs that are specifically related to gunshot injuries. This estimation can be considered as a supplementary component of the health care costs for a special interest. Using data provided by the Canadian Institute for Health Information (CIHI), we are able to estimate the medical treatment costs for gunshot injuries that were intentionally caused by other people. To be consistent with the scope of the present study, accidents and intentional self-harm such as suicides are excluded.

Unlike health care costs, many other costs are not directly observable as there is no monetary transaction. For example, in addition to hospitalization, victims might spend days recovering from physical injuries and emotional disorder, meeting with police, prosecutors or attending court proceedings. As a result, they may take time off from their daily activities and hence, there is a productivity lossto both victims and to society as a whole.

The productivity losses in this study account for both paid and unpaid production activities. The lost value of paid activities is straightforward to measure. For example, to estimate the value of lost wages, knowing how many days of absence and the average daily income would be sufficient. However, the productivity losses of unpaid work activities, such as searching for jobs, household services, raising children, caring for aging parents, shopping and volunteering, are more complicated to evaluate. Due to data limitations, we narrow the range of unpaid activities to three items, school attendance, household services and child care. For other types of activities, the value of household services is used as a proxy for estimation.Loss of missed school days is considered as a victim's forgone benefits as he/she may take additional time on his/her own or require extra lessons from private teachers or tutors to catch up with the class. The national daily tuition fee for undergraduate studies will be used. Lost household services and child care are valued according to the wage rate for a market substitute, for example, household workers or childcare service providers.

There are other types of productivity losses. People may feel angry, fearful and hurt, or may develop depression, anxiety and sleeping problems as a consequence of victimization. These negative impacts could result in a workplace disability or under-performance. People may feel it harder to concentrate on their work or take a longer time than usual to complete a project, which also may be considered productivity losses. Again, as we are lacking data, it is not possible to measure the magnitude of disappointment or the outcome of having 3 hours less sleep.

While tangible costs seem to be easily estimated, intangible costs, such as pain and suffering, are more difficult to quantify. Placing a monetary value on intangible items is subject to considerable uncertainty and controversy. There is now an extensive literature on the estimation of intangible costs and many methods have been explored such as willingness-to-pay (Ludwig and Cook, 2001; and Cohen et al. 2004) and jury awards (Cohen, 1988; Cohen et al., 1994; and Miller et al., 1996). However, no one has been able to produce estimates without much critique. Notwithstanding the differences in method among studies, the consensus remains that intangible costs are often the most expensive component of victim costs. In this study, the estimation for intangible costs will be conducted based on previous research.

It is important to consider the length of time over which the costs will be calculated. In this project, we try to capture both short and long-term effects of the victimization. Therefore, in addition to the current costs, costs occurring in the future as a result of the violence will also be examined. Long-term costs include items such as lifetime medical support services and lost future income due to the inability of performing job functions for the rest of the person's life. In the event of considering a future loss, all the relevant future costs will be computed into the present value for the year 2008. A summary of the categories for third party costs is included in Table 4.

Third-party Costs

Table 4: Categories of Third-Party Costs

  1. Funeral Service Expenses
  2. Loss of Affection/Enjoyment to Family Members
  3. Other People Harmed/Threatened
    • 3.1 Health Care
    • 3.2 Productivities Losses
  4. Social Service Operating Costs
    • 4.1 Transition Home/Shelter
    • 4.2 Crisis Lines
    • 4.4 Victim Services
  5. Other Related Expenditures
    • 5.1 Firearms Action Plan
    • 5.2 Investments to Combat the Criminal Use of Firearms

Crime has its most significant impact on victims, but others suffer as well. Family members may grieve the loss of a loved one or take time off from their daily activities to accompany victims (e.g., to court or doctor's appointments). In addition to the victim, there might be other people who get injured or threatened during the incidents. They might require medical attention or develop emotional disorders as well. Just like victims, all these consequences may cause underperformance of their work or reduce the quality of their lives. Furthermore, in response to this social illness, governments provide various victim services to help victims, and develop prevention programs, and other initiatives. For example, to enhance the capacity of law enforcement to fight gun crime, the federal government allocates approximately $50 million over five years for the Investments to Combat the Criminal Use of Firearms initiative which is aimed to improve the national collection, analysis and sharing of firearms-related intelligence and information. All these costs are reflected in the third-party costs.

Many people would argue that no amount of money would be adequate to compensate the victims or their families, especially in those fatal crime cases. This is true. Few people would voluntarily surrender their life for any amount of money. However, as Cohen (2005) says, “the cost of crime” is nothing more than the “benefit of reducing crime”. Therefore, the costs of a crime shall be interpreted in using this as guidance. For example, the costs of one firearm-related crime can be interpreted as the amount that would be saved if this incident is prevented. It can be considered as the amount that society would be willing to spend to prevent (reduce) one incident of firearm-related violence from occurring. These estimates are important numbers to have in a society where resources are scarce and where many programs - more police on the street, more funding to health and welfare, more programs to protect the environment, new highways or more parks - are always competing for tax revenue.

3.3 Data Sources

While the 2009 GSS is the main data source providing the prevalence and consequences of firearm-related violent crime, various other data sources are used as well. The major data sources are discussed as follows.

Uniform Crime Reporting Survey (UCR)

The UCR survey is an administrative survey that captures detailed information on crime reported to and substantiated by police. In this way, this survey is limited in that it does not capture personal victimizations that are not reported to police. There are several factors which can affect the proportion of crime that comes to the attention of police, including the willingness of the public to report crime or the way police report crime to the UCR. When estimating the criminal justice system costs, the UCR2 data, instead of the GSS data, will be used to provide the number of incidents for firearm-related violent crime since police-reported information, as compared to self-reported data, would better suggest the actual government resources that have been allocated to address the problem. Importantly, though, while the data collected by the GSS is limited to certain offences, the UCR collects information on all Criminal Code violations. Furthermore, the UCR2 micro data has a flag indicating the presence of firearm in each incident. In 2008, the national coverage of the UCR2 micro data was 98%.

Adult Criminal Court Survey (ACCS) and the Youth Court Survey (YCS)

The ACCS and the YCS provide statistical information on appearance, charges and cases in adult and youth criminal courts. Specifically, the surveys include information on court caseloads, characteristics of cases such as offence types and the average court elapsed time, and characteristics of persons moving through the court system such as gender/age, court outcomes, sentence types and sentence lengths imposed, etc. Both surveys have a flag indicating the presence of firearm in the case. Information from Quebec's municipal courts (which account for approximately one quarter (25%) of Criminal Code charges in that province) is not yet collected and reported to ACCS . According to Statistics Canada, this 25% amount should account for at least 5% of the total cases at the national level.

General Social Survey (GSS)

The GSS gathers data on social subjects in order to monitor changes in the living conditions and well-being of Canadians over time and provide immediate information on specific social policy issues of current or emerging interest. The GSS on Victimization is conducted every five years and asks Canadians aged 15 years and older about their personal experiences of criminal victimization, regardless of whether or not the incident was reported to police. The 2009 GSS , Cycle 23, is the fifth cycle dedicated to the topic of victimization. The purpose of this cycle is to better understand how Canadians perceive crime and the justice system and their experiences of victimization. It is the only national survey of self-reported victimization which provides data on criminal victimization for the provinces and territories.

As not all crimes are reported to police for various reasons, the GSS survey provides an important complement to the officially police-recorded crime statistics. Since the impact of victimizations (not reported to police) has no less of an effect on victims and society at large, using the self-reported data recorded by the GSS to estimate victim costs would capture a more fulsome and accurate reflection of the prevalence of firearm-related crimes in Canada, as apposed to the police-reported data.

In 2009, there were 19,422 respondents of the telephone interviews of Cycle 23. Each person represented roughly 1,400 people in the non-institutionalized Canadian population aged 15 years and over. Questions that provide information on the prevalence, nature and outcomes of criminal victimization include, but are not limited to: “Did you ever receive any medical attention at a hospital as a result of the violence”; “Did you stay in hospital overnight”; and “Did you ever have to take time off from your everyday activities because of what happened to you”, etc.. Particularly, the GSS has one important question indicating whether the criminal act was committed with a gun. The GSS also provides information on the characteristics of individual respondents such as gender, age, annual income, education, and main daily activities, which is useful to better identify and quantify the consequences of the violence.

National Ambulatory Care Reporting System (NACRS) and Discharge Abstract Database (DAD)

The two databases that are maintained by the Canadian Institute for Health Information (CIHI) provide valuable insight on volumes and types of cases presenting to Canadian hospitals. Both emergency and ambulatory care and hospital-based acute inpatient case are among the largest-volume patient activities in Canada, making them key components of health services in Canada.

The NACRS contains data for all hospital-based and community-based ambulatory care in Ontario: day surgery, outpatient clinics and emergency departments. The DAD records information on the acute inpatient hospitalizations and hospital discharges across Canada excluding Quebec. Detailed information includes type of ambulances used, types of firearm, average length of hospitalization, average treatment costs, number of patients that survived after treatment, discharge outcomes (home or transferred to other institutions) and demographic information such as gender and age of victims, etc. Note that both datasets do not have records from federal hospitals, e.g. prisons and veteran hospitals.

As a supplementary component, the CIHI data will be used to estimate the health care costs that are specifically related to gunshot injuries that were intentionally caused by other people. To be consistent with the scope of the present study, accidents and intentional self-harm such as suicides are excluded.

In summary, major data sources used in this project are:

  • Uniform Crime Survey 2 (UCR2), 2008
  • Police Administration Survey (PAS), 2008
  • Adult Criminal Court Survey (ACCS), 2007/08,2008/09
  • Youth Court Survey (YCS), 2007/08, 2008/09
  • Court Personnel and Expenditure Survey (CPES), 2002/03
  • Prosecutions Personnel and Expenditure Survey (PPES), 2002/03
  • Adult Correctional Services (ACS), 2007/08,2008/09
  • General Social Survey, Cycle 23, (GSS), 2009
  • Canadian Discharge Abstract Database, 2008/09
  • Hospital Morbidity Database, 2008/09
  • National Physician Database, 2008/09
  • Transition Home Survey (THS), 2003/04, 2005/06
  • Existing literature and government reports

3.4 Limitations

The most substantial limitation in conducting a costing analysis is a lack of comprehensive data. While there is a rich source of information available in Canada on many facets of crime impacts, there are also many data gaps that preclude a fulsome estimation of the associated financial costs. For example, the average number of follow-up calls made by distressed victims to crisis line is not known for certain. Since this data is not routinely collected by crisis lines, an average must be assumed based on qualitative data obtained from front line workers. This situation is not unique to Canada. Given the complexity of the topic and the wide-ranging impacts of crime, it is difficult for studies on crime costs in any country to be wholly accurate.

It should also be noted that the 2009 GSS does not provide sufficient information to allow distinctions to be made as to whether an outcome of the violence was due to a one-time incident or a series of repeat and chronic incidents. For example, a victim might be hospitalized once as a result of a one-time victimization and alternatively, and another victim might be hospitalized multiple times in a given year on account of suffering from chronic victimization. However, only one-time hospitalization is recorded by the GSS , even if a victim has been subject to repeat victimization and went to hospital more than once. Given the lack of clarity, the cost estimates presented in this report will be based on the number of victims, not the number of victimizations experienced. Furthermore, since injuries and psychological trauma could be worsened when associated with repeat victimization, for example in the case where violence occurs within intimate relationships, the true victim costs may be significantly underestimated.

Another limitation of the 2009 GSS is that the survey does not account for victimization in the three territories. Various challenges, including language difficulties and incomplete telephone service pose a significant constraint to the collection of data in the Yukon, the Northwest Territories and Nunavut.

In addition, even though the information regarding gunshot injuries provided by the CIHI data carried a lot of details, the results are far from comprehensive. First, the medical costs related to gunshot injuries were largely underestimated. Victims with less serious gunshot wounds might choose to see a physician or a nurse, or some victims may threaten a doctor to provide the necessary treatment. These victims, therefore, are not captured in the CIHI data which is hospital-based. Secondly, the CIHI information is only for one-time treatment. For survivors who face the prospect of a lengthy recovery period or a permanent disability, the CIHI data does not capture the medical expenses for follow-up appointments or life-time treatments at other health care facilities. Finally, the coverage of neither the NACRS nor the DAD is nationally representative. While the DAD data contains the information from all provinces with the exception of Quebec, the NACRS covers all the emergency departments in Ontario only. Therefore, certain assumptions and adjustments will be made during the estimation.

Finally, note that all survey data might be subject to both sampling error and non-sampling error such as non-response. This data quality issue is particularly pertinent to the 2009 GSS data. For example, people without language skills in either English or French were not able to respond to the telephone interviews; households without telephones or persons (e.g., young, single or urban Canadians) with cellular phone service only were also excluded from the survey. Therefore, somewhat different results might have been obtained if the entire population had been surveyed. Furthermore, wording of the questions and respondents' personal evaluations may also affect the answers given. As a self-reported survey, the answers are based on interviewee's memories, judgement and opinion, and in certain cases, people might not tell the truth for various reasons.

Considerations of firearm-related violent crime from an economic perspective reveal only one dimension of this complex social problem. Moreover, it is not always possible to account for all the consequences of the crime and hence, the estimates given in this study are far from complete. While direct effects are obvious and easy to estimate, indirect effects are much less obvious, and extremely complicated. The limitations discussed thus far are important to bear in mind when reviewing the report. Both the methodologies and the estimation can be revised and improved upon the arrival of new information and research in this area.

Date modified: