VI. Practical and Policy Implications of the Age–Crime Relationship
The debate surrounding the relationship between age and crime has also highlighted some practical and policy implications. Existing criminal justice policies have often been assessed in relation to the implications of the age– crime curve. For example, strategies such as the three strikes law (according to which courts are required to hand down a mandatory incarceration sentence to offenders who have been convicted of felonies three or more times) have been criticized in that, by the time the penalty for a third strike is implemented, the offender is likely at the end of his or her criminal career and would age out of criminal involvement regardless of the severity of the penalty. Other issues arise with regard to the appropriate crime reduction strategies implied by the age–crime curve and forecasting future trends in crime rates.
A. Targeting Participation Versus Frequency
The distinction between participation and frequency highlighted in the criminal careers debate proves to be an important consideration for crime policy. Hirschi and Gottfredson (1986) argued that programs targeted at reducing participation rates (i.e., reducing the proportion of the population that is engaging in criminal behavior) will have the largest effect on crime rates. They advocated for early intervention programs in particular. Farrington (1986) likewise suggested that, because the aggregate age–crime curve represents differences in participation, the best strategy to reduce crime is to prevent its onset by investing in early intervention programs. Blumstein and colleagues (1988), however, suggested that this is only one approach to decreasing crime. A second approach would be to reduce the frequency of offending among active offenders, which would involve more criminal justice strategy. These authors argued that there is a small group of offenders with a high frequency of offending and a relatively flat, stable age–crime curve (e.g., “chronic offenders” or “career criminals,” recognized as early as 1972 by Wolfgang, Figlio, & Sellin). A strategy such as selective incapacitation, which is targeted at reducing the frequency of offending among these chronic offenders, might be recommended. Selective incapacitation, however, relies on the assumption that these chronic offenders can be reliably identified before they are involved in an extensive number of offenses, something that has proven to be a difficult prospect (Gottfredson & Hirschi, 1986). Blumstein and colleagues did not dispute the difficulties in identifying these career criminals but suggested that they remain a valid topic of criminological inquiry.
B. Forecasting Crime Rates
Age has also become a major factor in explaining changes in crime rates over time and in forecasting future crime trends. For whatever theoretical reason, scholars have concluded that the age–crime curve reflects changes in the prevalence of offending among certain age groups. It is logical, then, that changing numbers of adolescents and young adults in the population should produce corresponding changes in crime rates (Phillips, 2006). This provides the potential opportunity to forecast changes in crime trends based on the age distribution of the population (Marvell & Moody, 1991).
Some research has suggested that the dramatically increasing crime rates during the 1960s and 1970s were attributable, at least in part, to demographic changes in the age structure of the population. Steffensmeier and Harer (1999) also suggested that the decline in crime during the first half of the 1980s was partly due to the declining population of teenagers. During this time, the sizable population of Baby Boomers was moving out of the most crime-prone years (i.e., aging out of crime). However, Marvell and Moody (1991) noted that although forecasts suggested a massive decline during the 1980s as the Baby Boomers aged out, the decline occurred for only the first half of the decade. Crime rates then increased again to record highs in the early 1990s even as the size of the teenage population declined (Fox, 1996; Marvell & Moody, 1991). Despite this confusion and apparent complexity in using age to predict crime rates, the declining crime rates during the 1990s were again attributed largely to the declining population of young adults (Steffensmeier & Harer, 1999).
In 1999, Steffensmeier and Harer found that changes in the age composition of the population did not appear to account for changes in crime rates as measured by both the Uniform Crime Reports and the National Crime Victimization Survey. Phillips’s (2006) cross-national research also finds no real relationship between the size of the youth population and crime rates. Most important, she found that the relationship between the percentage of young people in the population and homicide is attenuated when other criminogenic social conditions (e.g., low social control, high economic deprivation) are present. She suggested that the relationship between age and crime is complex and that the exact nature of the relationship depends on other social and cultural conditions. Levitt (1999) also concluded that, although forecasting crime rates on the basis of the number of teenagers in the population may be a logical assertion, the magnitude of the impact of age structure on crime remains unclear. His research suggests that changes in the age structure of the population account for no more than a 1% change in crime rates per year.
Marvell and Moody (1991) argued that demographic changes should not be used to forecast crime trends, because the age–crime relationship may not be strong enough to base predictions on, and other criminogenic forces may be more important. Despite the apparent complexity of this relationship, scholars and the popular media continue to forecast trends in crime rates based on the age structure of the population. In 1996, for example, Fox pointed to a “demographic time bomb” of crime and violence related to the increasing population of adolescents and young adults expected through 2010. This echoed earlier suggestions that a new crime wave would be fueled by a new, large generation of “super-predators” (Steffensmeier & Harer, 1999). By 2005, the teenage population was expected to reach its largest size in three decades (Fox, 1996). However, crime rates declined substantially throughout the decade of the 1990s and remained low in the early 2000s (Steffensmeier & Harer, 1999). Marvell and Moody (1991) summarized this difficulty by concluding from their review of 90 studies that “the age/crime relationship is far from established” (p. 251), limiting its utility in predicting future crime trends based on the age distribution of the population.