- 1 This research was supported by a grant from SSHRC. I would like to thank anonymous reviewers as wel (...)
1Crime in Canada, as recorded by official statistics, increased annually during the 1960s and 1970s1. This led to scepticism among many scholars regarding the validity of the crime figures. According to Hagan (1991), the growth in crime during that period can be explained partly by better recording practices and by a greater police enforcement of minor statutes. Despite distortions caused by these changes, Hagan concludes that there has been a real increase during that period in crime and violence, as well as in rates of alcohol and drug abuse. In the 1980s, official statistics for some types of offences show periods of stability and even slight decline, but trends for violence continued to grow sharply. The 1990s show a strikingly different picture. Since 1992, most categories of crimes have declined significantly in Canada.
2In “The Crime Drop in America” (Blumstein and Wallman, 2000), experts examine the unprecedented fall of crime in the USA throughout the 1990s which can be observed both in official crime statistics and the national crime survey (Rennison, 2000). It is interesting to note that no one had foreseen the amplitude of the crime drop before it happened. Our collective inability to foresee future crime trends, even with the use of sophisticated time-series analytical models, demonstrates that criminological macro-theories have yet to be developed. The reality of the decline of crime has only been acknowledged in recent years and is now attracting research. Chaiken concludes, after reviewing a Statistics Canada publication, that ‘the longer pattern in Canada is essentially flat’ (2000, p.1). This is troublesome because all data indicate that a drop in crime also occurred in Canada. Although Chaiken does not acknowledge that fact, drops in crime during the 90s have also been observed elsewhere. In Germany, the total offence rate dropped by 7.8% between 1993 ands 1999 (PKS Berichtsjahr, 2001). In England, the total number of crimes recorded by the British Crime Survey dropped by 15.6% between 1993 and 1999 (Povey, 2001).
3In this paper, long term crime trends in Canada and the United States are examined in order to establish the concordance between the two countries. Then, a detailed analysis of the evolution of crime between 1993 and 1999 is presented. Finally, factors that may help explain recent trends are examined.
- 2 For detailed statistics, see www.canadianambassy.org (...)
4Canada is the country that most resembles the U.S. First, more than 90% of the 30 million Canadians live within 100 miles from the U.S. border. Second, our history of colonisation is generally comparable, although Canada has retained a more continental influence than the U.S. (Lipset, 1990). Third, both countries have historically been lands of immigration from around the world. Perhaps more important today is the fact that our popular cultural world is mostly undifferentiated, whether in movies, sports or television (Quebec being the exception). With the trend towards globalization and particularly the NAFTA agreement, exchanges between our countries have increased rapidly in the last decade. Canada is by far the largest trading partner of the U.S. (and vice versa) in both imports and exports, with a total trade figure of $434 billion in 1999 as compared to $284b with Japan, $251b with the U.K. and $238b for Mexico2. Our commercial exchanges amount to more than one billion U.S. dollars a day. There are, however, major differences between Canada and the United States. Canada is a more egalitarian society. This is most visible in easy access to education and in universal access to health care services. There are also less visible differences, such as differences in personal values. According to Lipset (1990), Canadians are more tolerant toward others and more open to different races, religions or sexual orientations.
5In terms of crime incidence, there are similarities and differences between Canada and the United States. Mayhew and van Dijk (1997), who conducted an international crime victimization survey, found that in 1995 the overall national victimization risk was 24.2% for Americans and 25.2% for Canadians. The prevalence of most types of crimes is comparable even for crimes such as robbery (1.3% for Americans vs. 1.2% for Canadians). Two factors stand out in an analysis of the crime problem: the greater lethality of interpersonal conflicts in the U.S. because of easy access to firearms, and the disproportionate crime problem in large U.S. cities (Ouimet, 1999).
6In terms of crime control, Canada and the U.S. are very different. In the U.S., there are more per capita police officers, courts are more likely to sentence offenders to jail or prison, average terms of imprisonment are longer in the U.S, and possibilities for parole are more likely in Canada. All these factors, in association with high rates of violence, contribute to the fact that the incarceration rate in the U.S. is more than four times greater than in Canada.
- 3 For Canada, data for 1901-1961 was obtained from Buckley (1963) and Leacy (1983), while data for th (...)
7In order to compare the general evolution of crime trends in both countries, long-term trends in the homicide rate were examined. This offence was chosen because data are available for long term periods, it is an accurately measured variable, shows little differences in definitions over time or space and is correlated with the incidence of less serious forms of crime. Figure 1 displays the Canadian and American homicide rate from 1901 to 1999.3
Figure 1: Trends in homicide rate per 100,000 for Canada and the United States from 1901 to 1999 (See appendix)
- 4 In order to test the presence of a trend in the ratio, a regression was performed with T as the ind (...)
8Although the homicide rate is three to four times higher in the U.S. than in Canada, trends are comparable between both countries. Figure 1 shows that in both countries, homicide rates increased in the three first decades of the century, culminating in the early thirties, a period of economic depression. Rates dropped considerably from the mid-thirties to the fifties, despite a small post-war jump. In both countries, the homicide rate increased dramatically between the early sixties and the late seventies before reaching a plateau in the early nineties, after which it decreased again. Homicide rates for both countries in 1999 are comparable to rates during the mid-sixties. The concordance in the general trends of both countries’ homicide rates is remarkable. As one can imagine, the ratio between both curves is flat from 1962 to 1999 (it averages 3.6).4 Such a result also rejects the hypothesis of a convergence between the U.S. and Canada that was proposed by Lenton (1989) and correctly rejected by Hagan (1991). The long-term parallelism between homicide trends in both countries can be interpreted as an indication that the forces behind these trends are very general and diffused across North America.
9Many factors may have contributed to the last decade’s decline in crime rates. In a recent book, editors Alfred Blumstein and Joel Wallman (2000) asked contributors to discuss the changes in specific areas that might have played a role in the downward trend of violence. In chapter two, Blumstein shows that the homicide peak of the early nineties is mostly attributable to an epidemic of young urban black handgun homicides. Yet, the homicide rate for whites, older people or non-urban people has also gone down in the 90s. There are three major explanations for the drop in crime that are examined in the book: demographic trends, incarceration policies and police practices.
10The demographic approach is to look at downward trends in crime as a consequence of the diminishing number of people in the crime-prone age category, namely teenagers and young adults. Steffensmeier and Harer (1999) argue that changes in age composition made only a small contribution to the downward trend of crime in the years of the Clinton administration. There is scepticism over the demographic explanation because the impact of changes in the population structure on crime should, for many observers, have taken place earlier. The number of teenagers and young adults peaked around 1980 and dropped during the 80s. The explanation for the lag between the two phenomena may well be the presence of a cohort effect. The cohort effect would specify that people born during the early 60s showed higher levels of crime involvement than any cohort before or after, and were criminally active for longer periods than other cohorts (in large part explainable by difficulties integrating the job market during the 80s). Therefore, the impact of the diminishing number of youths has been delayed, and this explains the temporary lag between the two phenomena.
11The role of increases in the incarceration rate on the crime rate was carefully examined by Spelman (2000). He estimates the relationship between incarceration and crime and comes up with an elasticity measure of –0.4, which means that an increase in incarceration of one percent will be followed by a drop in crime of 0.4%, or 4/1000. For Spelman, about one quarter of the drop in crime during the 90s can be explained by increases in incarceration. However, not all scholars share the view that the use of incarceration is related to trends in the crime rate. An analysis provided by the Justice Policy Institute (2000) shows that the State of Texas increased its prison population throughout the 90s to reach an incarceration rate of 1,035 per 100,000 residents in 1999, second only to Louisiana, while its drop in crime for the 1995-1998 period was 5%, far behind the drop of 23% in California or 21% in New York.
- 5 The unstandardized error term of the equation [Y(1993 incarceration rate) = a + b (1993 homicide r (...)
12Our own calculations of the correlation between the 1993-1998 percentage change in homicide rates and incarceration rates for 48 states (DC, Hawaii and Alaska not included ) stands at 0,085 and is not statistically different from zero (correlations for other index crimes are all positive but non-significant at a 0,05 level). In other words, trends in states’ incarceration levels are not associated to trends in crime levels. However, as explained elsewhere (Ouimet and Tremblay, 1996), the error around the correlation between the 1993 murder rate and the incarceration rate (correlation of 0,768 in 1993) is strongly and positively related to the percent change in the incarceration level between 1993 and 1998, suggesting that American states are adjusting their incarceration levels to the states’ norm given their crime rate.5 For example, the state of Delaware had an incarceration rate in 1993 of 394 (which was more than the rate of 226 which it should have had in light of its murder rate), “only” increased its incarceration by 8,8% between 1993 and 1998. At the other extreme, West Virginia, with an incarceration level of 98 in 1993 (much lower than expected) has managed to almost double its incarceration rate to 192 in 1998. Ouimet and Tremblay (1996) offer a political model for explaining these variations in the incarceration rate.
13Another explanation for the drop in crime concerns policing. Eck and Maguire (2000) examined the possible impact of changes in policing on the drop in crime, but are generally quite critical of the claims that have been made by police officials about their role in the general downward trend in crime. The main problem with the policing explanation is that innovative police practices such as gun patrols (stop and frisk) or Compstat type systems had been implemented after the crime rate had already begun declining. Moroever, the rate of crime dropped even in cities that had not experienced major changes in policing. In sum, we know too little about the specific impact of different police strategies such as problem-oriented policing, community policing or zero-tolerance policing to allow us to evaluate the causal relationship between policing and crime trends.
14Other explanations have been advanced to explain the drop in crime. While Rosenfeld (2000) acknowledges the impact of demographic and incarceration factors, he argues that other variables have to be considered to explain the downward trend in homicide, one of which might be the “civilizing process” described by Elias (1939). Elias shows how people gradually become more sensitive to violence and are therefore gradually expressing greater intolerance towards such behaviors. Holloway (1999) advances an explanation developed by Levitt and Donohue that can account for diminishing crime rates in most places in the U.S. during the 90s (whether policing practices have changed or not, whether state imprisonment rates changed much or a little, etc). They argue that increases in abortion during the late 70s and 80s resulted in the reduction of the numbers of unwanted children that would have growne to become delinquents. What the authors do not tell us is that the number of single-parent families also increased during the late 70’s and 80’s, and this should have contributed to a growth in juvenile delinquents during the 90s. Another explanation was proposed by Felson (1998), who concludes that the decline in the use of paper money bills has contributed to a general reduction in theft, robbery and even homicide. It is not clear, however, why this trend would have had any impact on the rates crimes such as car theft or burglary.
15The contributing authors in “The Crime Drop in America” focus on serious violence, namely homicide and robbery. Sexual assault and aggravated assault are not examined because it is believed that UCR statistics are not reliable for these behaviours. Little mention is made of the fact that very common crimes such as burglary and car theft are also declining. The result of this selection by the authors is to look for specific factors, rather than those that may explain the many varieties of trends in crime. For example, if one concentrates on homicide, the question of guns becomes central, but guns play little role explaining rates of burglary or car theft. Second, no effort is made to comparatively sort out the different possible causes of the decline.
- 6 Canada’s three territories, Yukon, the North West Territories and Nunavut were not included in the (...)
16The year 1991 can be considered as the peak in the upward trend for many types of crimes in both Canada and the US. In the eight years that followed, the volume of several types of crime decreased annually. To study the Canadian context, five regions were constructed out of ten province; Prince Edward Island, New Brunswick, Nova Scotia and Newfoundland were grouped into the “Atlantic region”, while Alberta, Saskatchewan and Manitoba were grouped in the “Prairies region”. This has been done to provide an easier analysis of the trends in crime and because penitentiary data are only available on the basis of those regions.6
17To study crime, we selected a few well-defined offences that are often used in trend analysis and represent the principal crimes that result in the incarceration of the offender. Canadian data were coded from Statistics Canada’s CANSIM data sets, while American data were extracted from the Bureau of Justice Statistics website. In terms of definitions for homicide, robbery, burglary and motor vehicle theft, the rates can be compared directly between both countries because the statutory definitions of these offences are fairly similar. The rates for sexual assault and assault should not be compared across countries since the American rape and aggravated assault concepts are more limited than in Canada (there is no way to isolate aggravated assault and aggravated sexual assault in Canadian statistics). Table 1 presents the crime data for Canada and the United States in 1991 and 1999.
Table 1: Rate per 100,000 population for selected offences for Canadian regions and the USA in 1991 and 1999
66Motor vehicle theft
752 330 800
766 845 700
779 914 200
784 609 500
793 218 500
8026 918 700
122Motor vehicle theft
1312 368 374
1327 333 283
1345 078 177
1354 009 922
13630 201 303
179Motor vehicle theft
194Table 1 shows that there has been an important reduction in crime rates for most categories of crime in most regions of Canada. The drop varies from 23% (assault and robbery) to 43% for homicide. The homicide rate has declined in all regions; the reduction ranges from 35% to 58%. With the exception of motor vehicle theft, the frequency of all crimes decreased in all regions except robbery in the Prairies (no change). The only crime without a declining rate is motor vehicle theft, which nevertheless declines on a per 100,000 registered motor vehicles basis. Table 1 also reports crime figures for the U.S. for the years 1991 and 1999. It is difficult to determine in which country crime has decreased the most. In fact, our interpretation of the data is that crime has decreased at the same rate in both countries.
195The data show that the drop in crime occurred in Canada as well as in the United States, and that the trends are comparable for most types of crime. This would indicate that the causes of the decline cannot be found in very specific factors that may have played a role in some jurisdictions. In this sense, searching for causes of the decline in specific forms of crime – such as the drop in youth firearm- related homicides – is misleading.
196A number of factors may have played a role in the explanation of the recent trend in crime in Canada. We grouped them in two dimensions; endogenous or exogenous. Endogenous factors are those that are part of the criminal justice system and over which agencies have some control. Exogenous factors are those which exist outside the scope of the criminal justice system.
- 7 Data on unemployment demography comes from www.fedstats.org and www.statscan.ca. (...)
197Endogenous factors: In Canada during the 1990s, there has been no sign of toughening in various sectors of the criminal justice system. In fact, given the difficult economic situation of the different governments during this period, federal and provincial governments were disinclined to develop initiatives that would increase government spending. Debt reduction was the main objective for most governments. Statistics Canada provides data to substantiate this claim: spending related to justice in Canada for 1992-93 and for 1996-97 were as follows: $5,717 vs. $5,856 millions for police, $867 vs. $857 millions for criminal courts, and $1,880 vs. $1,969 millions for adult correctional services. In fact, the growth in the budget of justice- related organizations was below the level of inflation. Table 2 presents information relating to policing and correctional sectors for both the U.S. and Canada in 1991 and 1999. For comparative purposes, the incarceration rate is based on the number of adults incarcerated in relation to the total population, not on the officially published adult incarceration rate per 100,000 adults, which stands at 140 in 1999 (Thomas, 2000). 7
Table 2: Endogenous and exogenous indicators in USA and Canada for 1991 and 1999
206Population in millions
213Rate of police officers per 1000 (numbers in parentheses)
226Incarceration rate per 100,000
227(numbers in parentheses)
240Global unemployment rate
247Proportion of the population aged 20 to 34 (1990 and 1999)
- 8 Passed in 1991, Bill 147 in Québec allows the prison warden to transfer to community correctional (...)
- 9 This new disposition was created by the Sentencing Reform Bill (C-41) that was proclaimed in 1996.
254Table 2 shows that in Canada both the number of police officers per 1,000 and the incarceration rate fell, respectively, by 11% and 3%. In the US, increases have been observed in both the number of police officers per 100,000 (+11%) and in the incarceration rate (+42%). The picture that emerges from both countries is therefore dramatically different. The fact that neither police nor correctional indicators show an increase in Canada during the nineties does not mean that no important changes have occured within these sectors. For example, police organizations improved different aspects of their work throughout the nineties, including a greater use of computers and databases and the creation of specialized units or task forces. This was also true in the U.S. On the correctional side, agencies responded to increases in the number of people sent to prison in the early nineties by allowing early release for non-violent offenders. In Quebec, for example, since the early 90s, non-violent offenders have been expected to be put under community correctional supervision after serving one-sixth of their sentence (Corbo, 2001).8 Even courts help relieve the pressure by employing a new alternative form of imprisonment called a conditional sentence, which is a prison sentence that is served in the community.9 Although not necessarily well understood or popular within the general population, these measures made it possible, in the early 1990s, to increase the number of people sentenced by courts and the number of people sent to prison without significantly increasing the average size of the incarcerated population. Another aspect relating to endogenous factors is the question of drug crimes. In Canada, there has simply not been a “War on drugs”. On the other hand, this “war” has greatly contributed to the incarceration trends in the U.S. in the 1980s and 1990s (Blumstein and Beck, 1999).
255Exogenous factors: There are two major social forces that may have played a role in explaining the declining rate of crime in Canada during the 1990s: demography and economic prosperity. The “baby boom” has had a more important impact in Canada than in any other industrialized country (Foote, 2000). The after-war boom continued into the early sixties, at which point the number of births dropped until the early seventies and has remained low since. Nineteen sixty two is the single year in which the greatest percentage of canadians were born (Foote, 2000). The large cohort born around 1960 became young adults at the turn of the eighties, when the recession arrived. During the 1980’s, there were hiring freezes, shutdowns and high unemployment rates. In 1992, the large cohort born around 1962 reached 30 years of age and in 1999 people in this category were approaching 40. The traditional age-crime curve tends to point to 18 years of age as the most violence-prone age, while the average serious offender is older (Farrington, 1986). Murderers, bank robbers and rapists are closer to 30 than to 20.
256Therefore, if one concedes that serious crime is not only a teenager or very young adult problem (which is obvious for sexual assault and domestic violence), the downward trend in crime observed in Canada might very well be explained by a progressive decrease in high crime groups, such as the 18-24, 25-29 and 30-34 age groups. Hence, the crime trends in Canada are compatible with the demographic explanation; property crimes (highly associated with teenagers and young adults) levelled off during the 80’s, while violent crimes (mostly committed by youths but also middle aged adults) decreased in the 90s. Data presented in Table 2 show that the number of people aged between 20 and 34 dropped by 18% in the U.S. and Canada between 1990 and 1999.
257The other exogenous factor involves the economy. Canada had very high unemployment figures throughout the 1980s. The situation improved in the early 1990s and continued to improve thereafter. Table 2 shows that the total unemployment rate dropped by 36% in the U.S. and by 27% in Canada between 1991 and 1999. Employment is now more easily available for young adults, especially in the world of service oriented part-time jobs. More stable and rewarding jobs and opportunities have grown during the 1990s as a result of the massive retirement wave that took place amongst early baby-boomers. The economic situation for young adults has therefore improved greatly during the 1990s, explaining at least a part of the drop (Freeman, 2000).
258Besides demographic and economic changes, other factors may play a role in explaining decreasing crime trends in Canada and in the United States. One of those factors might be a growing proportion of young adults pursuing higher education. According to Statistics Canada (2000): “In 1990, 20% of people aged 25 to 29 in Canada had less than high school education. By 1998, that percentage had dropped to 13%. Also, between 1990 and 1998, the percentage of individuals in this age group who had university degrees rose from 17% to 26%.” Another factor explaining the decline in the crime rate might be a decrease in public consumption of alcohol, presumably because of more stringent enforcement of DWI laws. In fact, the per-capita sale of spirits and beer in Canada has decreased slightly during the 90s (Statistics Canada, 1999), but public consumption in restaurants and bars might have dropped even more rapidly. These two hypotheses, as well as others, have to be examined more closely in further research.
260According to LaFree (1998, 1999), “The United States in the 1990s has experienced the greatest sustained decline in violent crimes since World War II”. This decline applies to most types of street crimes. Crime trends in Canada are very similar to those observed in the U.S. The quest for a general explanation should therefore focus on changes that have affected both countries, in all regions and for most forms of crime. If changes in the use of incarceration is to be invoked as an explanation, it would have to be shown why it worked in the U.S. but not in Canada. In terms of policing, contrary to the U.S., Canada has not increased the pro rata number of police officers. Nevertheless, advances in technology have characterized many police organisations during the 90s as well as changes in police operations. While police organizations are doubtlessly more efficient in 1999 than in 1991, we do not know what changes, if any, have had an impact on crime (Eck and Maguire, 2000). In Canada, there has been no movement toward more aggressive policing as was observed in many U.S. cities. As we have seen, the areas that are the most promising for explaining the crime drop in both countries have to do with demography and economic prosperity. However it is also possible that the two nations, as well as others, may have gone through a period of important changes in collective values.
261Collective values are the factors that are the most difficult to grasp for researchers. Nevertheless, they may play an important role in explaining recent crime trends. The work of Elias (1939), who studied the evolution of good manners across centuries, provides an understanding of how violent behaviour was progressively identified as “bad behaviour”. In the early middle Ages, knights took pleasure in torturing and killing peasants. As society evolved, this behaviour was no longer accepted. The process proposed by Elias seems quite straightforward: changes in social structures are accompanied by changes in people’s sensibilities. Identifying what in our society changed during the 80’s and 90s to cause our values regarding crime and violence to evolve so drastically would allow us to explain how and why crime rates decline.
262Homicide rates were high in the first three decades of this century. It might very well be that industrialization, massive immigration and poverty were factors that disrupted family structure and the social fabric. At mid-century, a new ethos emerged in our society, with strong families, disciplinary schools and religious fervour. Self-restraint was tantamount to good manners. The cultural revolution of the 60’s and 70’s, with liberalization and self-expression becoming the dominant ethos -- or pathos – has profoundly changed our society (Wilson and Herrnstein, 1985; Freeman, 2000). Crime rates, but also the rates of automobile accidents and suicides, increased rapidly. The late 1980s and 1990s, however, can be characterized by the progressive integration of a new ethos of moderation in drinking, drug use, sexual activity and even tobacco use. Many behaviours that were seen as acceptable or were not the object of public outrage only twenty years ago are now gradually integrated into our moral conditioning.