Abstract: This study extends research on cohort effects and crime by considering how bifurcated population dynamics and institutional constraints explain variation in homicide rates across racial groups in the United States. Drawing upon the extensive research on racial residential segregation and institutional segmentation, we theorize how the criminogenic influences of cohort characteristics elucidated in prior work will be greater for Black cohorts than for White cohorts. We assess our hypothesis by estimating Age-Period-Cohort Characteristic models with data for the total population and separately for the Black and White populations over the 1975–2014 period. The results reveal persistent effects of relative cohort size and nonmarital births on Black cohort-specific homicide rates but null effects among the White population. These effects follow Black birth cohorts across the life course, leading to higher rates of both homicide arrest and homicide victimization.
Summary and Discussion
Building on the seminal work of Richard Easterlin (1987), this study provides a
novel extension of the empirical literature on population dynamics and criminal
homicide. Consistent with the Easterlin thesis, our research is guided by the
assumption that large birth cohorts strain the regulatory capacity of social
institutions (e.g., families and schools), decrease informal social control, amplify
cross-cohort socialization, and increase competition for entry-level positions,
ultimately contributing to higher rates of cohort-specific violence. We depart
from prior scholarship by arguing that the study of cohort effects remains incomplete because scholars have yet to consider how the pernicious consequences
of residential segregation and labor segmentation entrenched within American
society have conditioned the influence of cohort characteristics on homicide
rates. Insofar as the key mechanism linking relative cohort size to criminal
conduct is the ability of social institutions to effectively integrate large birth
cohorts, it follows that the strongest cohort effects should operate within racial
groups over time. As such, the present study examined the relationships among
relative cohort size, nonmarital births, and age-by-race specific rates of homicide
for the years spanning 1975–2014.
The results from the empirical models generally support our racially bifurcated
perspective on cohort characteristics and crime. We find consistent evidence that
the proportion of nonmarital childbirths is positively associated with overall
rates of homicide arrest and victimization but no evidence of an effect of
relative cohort size on overall cohort-specific homicide rates over the past four
decades. As we elaborate in greater detail below, the discrepancy between our
findings and some prior research (e.g., O’Brien et al. 1999; Savolainen 2000) is
likely attributed to necessary differences in data source and observation period
between our work and others. When we turn to the APCC models separated by
race, the results are striking. The proportion of nonmarital births and relative
cohort size exert strong, positive effects on cohort-specific homicide arrest and
victimization rates among Blacks. The highest rates of homicide victimization
and arrest are observed among Blacks born during periods of relatively high
fertility and those born during times of high nonmarital births. We interpret
these findings in line with our key theoretical argument—the effects of cohort
composition on criminal homicide reveal themselves most strongly among the
Black population. As evidenced by the discrepancy in findings between the first
and third models in Tables 2 and 3, focusing on aggregated cohort effects for
the total population obscures important nuances in the racialized nature of the
influence of population dynamics on homicide trends in recent decades.
We speculate that the most likely culprit driving these differential effects is
a labor market that allows large White birth cohorts to edge Blacks out of
low-skilled positions. Such a mechanism does not require a far stretch of the
imagination. All else equal, labor market shocks, such as those associated with
large birth cohorts, portend that a large number of young adults will be vying
for a proportionately fewer of entry-level positions. When labor supply exceeds
demand, employers can be more discriminating in staffing decisions. Given the
storied history of discriminatory hiring practices in the United States and a labor
market clearly differentiated by race, it seems reasonable to expect that Blacks
will be hit especially hard during times of labor surplus. From this vantage
point, Black Americans may find themselves at a disadvantage when they are
born during a time of high fertility because they will encounter greater levels of
competition with other young African Americans and greater competition with
Whites who may edge into traditionally segmented positions. The deleterious
effects of cohort size follow Black birth cohorts across the life course, translating
into elevated rates of homicide victimization and arrest. This finding helps shed
light on the lack of an association between relative White cohort size and
homicide rates—because social institutions are better able to accommodate large
White cohorts by further disadvantaging Blacks, the otherwise criminogenic
influence of cohort size is mitigated among the White population.
Importantly, the relative cohort size findings are robust in the presence of
race-specific nonmarital fertility, which emerges as an independent predictor
of homicide for the age-specific total and Black crime rates, but not for agespecific White crime rates. This suggests that differences in the impact of relative
cohort size on race-specific crime rates cannot be attributed to differences in
supervision that arise from disparate nonmarital fertility trends. Additionally, it
hints at further ways in which the supervision capacities of institutions may be
more adaptive for White cohorts compared to Black cohorts.
To be clear, our intention is not to refute the rich scholarship stressing the
importance of neighborhood deprivation, subcultural norms, or persistent structural inequalities that are often invoked to explain differences in violence across
racial and ethnic groups. Instead, we hope to illuminate an often overlooked
demographic component of violent crime trends. Rather than supplant prior
explanations, we adopt the view that the racialized nature of cohort effects on
crime complements contemporary thinking on criminal violence. The inveterate
legacy of segregation and discrimination in American society has generated
vastly different social institutions clearly delineated along racial lines (Peterson
and Krivo 2010). Insofar as the mechanisms linking cohort characteristics to
criminal violence involve the capacity of such institutions to assimilate successive
generations, it follows that generations of Blacks and Whites have been raised
by families living in segregated neighborhoods, attended segregated schools, and
entered into segmented labor markets. The ability of such segregated institutions
to effectively socialize large cohorts of children is a defining factor in the
perpetuation of social inequalities, which have generated profound differences
in homicide rates across racial groups in the United States.
We would be remiss not to acknowledge several caveats with our analyses
that limit our ability to engage more directly with prior scholarship in this
area. For one, our empirical models necessitate age-by-race specific measures
of homicide and victimization. The SHR provides the longest-running source of
this information, but these data only extend back as far as 1975. While there
are clear advantages to using the SHR, our models do not directly align with
previous research (which has often relied on measures gleaned from the UCR
arrest records). Accordingly, the discrepancies between our findings for total
homicide offending and prior work may be due to broader issues with official
data collection over time. Because our analyses span different time frames,
we speculate that the differences in results for total homicide offending rates
may arise from differences in arrest reporting and subsequent disaggregation.
Regardless, our goal was not to dispute the pioneering work of others but rather
to explore more nuanced pathways through which race, population dynamics,
and segregation have influenced trends in criminal homicides in the United
States.
Relatedly, a growing body of literature demonstrates how exogenous shocks
and large-scale social changes can influence long-term crime trends (Rosenfeld
2018; Baumer, Vélez, and Rosenfeld 2018). When considered in the context
of the current study, we might anticipate such period effects to influence not
only homicide rates but to indirectly contribute to cohort characteristics by
differentially influencing the relative size and rates of nonmarital birth characterizing Black and White cohorts overtime. The most obvious examples are
the crack cocaine epidemic and the differential impact of mass incarceration on
communities of color. Indeed, both mass incarceration and the homicide boom of
the late 1980s had a disproportionate effect on specific cohorts of young, Black
men. While beyond the purview of the current study, it is entirely possible that
these period effects reshaped the long-term life chances of this cohort, far beyond
the historical moment which such events occurred.
We are also limited by our inability to examine the more proximate mechanisms linking cohort characteristics and criminal homicide, such as indicators
of school crowding, educational attainment, or labor market outcomes. And
indeed, the threats posed by omitted variable bias remain a considerable hurdle
in APCC models (O’Brien 2014). The most glaring omission in this regard is our
lack of measures of age-by-race specific unemployment rates, which would allow
us to examine directly whether labor market edging explains the strong effect of
cohort size on Black homicide rates. To our knowledge, no such measures exist
for the full period of observation used in our analyses. Rather than view this as
a critical limitation, we echo Robert K. Merton (1987) who argued that “before
one proceeds to explain or to interpret a phenomenon, it is advisable to establish
that the phenomenon actually exists, that it is enough of a regularity to require
and to allow explanation” (3). We view the empirical contributions described
here as a necessary first step toward confirming the differential impact of
population dynamics on violence within racial groups, thus laying the foundation
for further research into the mechanisms driving the disproportionate influence
of cohort characteristics on criminal homicide over the past 60 years.
Despite these caveats, our findings reaffirm the importance of systematically
incorporating demographic processes into criminological research. Social control
theories have long extolled the central role played by social institutions in
suppressing violent crime, but these theories have devoted little attention to the
ways in which rapid population growth might strain institutional capacity. Our
work further underscores the inexorable linkages between population dynamics
and institutional constraints in the propagation of racial inequality in the United
States. To paraphrase Richard Easterlin, year of birth indeed marks a generation
for life. In the context of criminal homicide in the United States, it is clear
that the enduring consequences of cohort characteristics for homicide offending
and victimization unfolded differently depending on race. Consistent with a
growing body of scholarship, the results presented here suggest that crime,
violence, and the perpetuation of racial inequality in the United States can
be best viewed through a historicized understanding of bifurcated population
dynamics.
Building on the seminal work of Richard Easterlin (1987), this study provides a
novel extension of the empirical literature on population dynamics and criminal
homicide. Consistent with the Easterlin thesis, our research is guided by the
assumption that large birth cohorts strain the regulatory capacity of social
institutions (e.g., families and schools), decrease informal social control, amplify
cross-cohort socialization, and increase competition for entry-level positions,
ultimately contributing to higher rates of cohort-specific violence. We depart
from prior scholarship by arguing that the study of cohort effects remains incomplete because scholars have yet to consider how the pernicious consequences
of residential segregation and labor segmentation entrenched within American
society have conditioned the influence of cohort characteristics on homicide
rates. Insofar as the key mechanism linking relative cohort size to criminal
conduct is the ability of social institutions to effectively integrate large birth
cohorts, it follows that the strongest cohort effects should operate within racial
groups over time. As such, the present study examined the relationships among
relative cohort size, nonmarital births, and age-by-race specific rates of homicide
for the years spanning 1975–2014.
The results from the empirical models generally support our racially bifurcated
perspective on cohort characteristics and crime. We find consistent evidence that
the proportion of nonmarital childbirths is positively associated with overall
rates of homicide arrest and victimization but no evidence of an effect of
relative cohort size on overall cohort-specific homicide rates over the past four
decades. As we elaborate in greater detail below, the discrepancy between our
findings and some prior research (e.g., O’Brien et al. 1999; Savolainen 2000) is
likely attributed to necessary differences in data source and observation period
between our work and others. When we turn to the APCC models separated by
race, the results are striking. The proportion of nonmarital births and relative
cohort size exert strong, positive effects on cohort-specific homicide arrest and
victimization rates among Blacks. The highest rates of homicide victimization
and arrest are observed among Blacks born during periods of relatively high
fertility and those born during times of high nonmarital births. We interpret
these findings in line with our key theoretical argument—the effects of cohort
composition on criminal homicide reveal themselves most strongly among the
Black population. As evidenced by the discrepancy in findings between the first
and third models in Tables 2 and 3, focusing on aggregated cohort effects for
the total population obscures important nuances in the racialized nature of the
influence of population dynamics on homicide trends in recent decades.
We speculate that the most likely culprit driving these differential effects is
a labor market that allows large White birth cohorts to edge Blacks out of
low-skilled positions. Such a mechanism does not require a far stretch of the
imagination. All else equal, labor market shocks, such as those associated with
large birth cohorts, portend that a large number of young adults will be vying
for a proportionately fewer of entry-level positions. When labor supply exceeds
demand, employers can be more discriminating in staffing decisions. Given the
storied history of discriminatory hiring practices in the United States and a labor
market clearly differentiated by race, it seems reasonable to expect that Blacks
will be hit especially hard during times of labor surplus. From this vantage
point, Black Americans may find themselves at a disadvantage when they are
born during a time of high fertility because they will encounter greater levels of
competition with other young African Americans and greater competition with
Whites who may edge into traditionally segmented positions. The deleterious
effects of cohort size follow Black birth cohorts across the life course, translating
into elevated rates of homicide victimization and arrest. This finding helps shed
light on the lack of an association between relative White cohort size and
homicide rates—because social institutions are better able to accommodate large
White cohorts by further disadvantaging Blacks, the otherwise criminogenic
influence of cohort size is mitigated among the White population.
Importantly, the relative cohort size findings are robust in the presence of
race-specific nonmarital fertility, which emerges as an independent predictor
of homicide for the age-specific total and Black crime rates, but not for agespecific White crime rates. This suggests that differences in the impact of relative
cohort size on race-specific crime rates cannot be attributed to differences in
supervision that arise from disparate nonmarital fertility trends. Additionally, it
hints at further ways in which the supervision capacities of institutions may be
more adaptive for White cohorts compared to Black cohorts.
To be clear, our intention is not to refute the rich scholarship stressing the
importance of neighborhood deprivation, subcultural norms, or persistent structural inequalities that are often invoked to explain differences in violence across
racial and ethnic groups. Instead, we hope to illuminate an often overlooked
demographic component of violent crime trends. Rather than supplant prior
explanations, we adopt the view that the racialized nature of cohort effects on
crime complements contemporary thinking on criminal violence. The inveterate
legacy of segregation and discrimination in American society has generated
vastly different social institutions clearly delineated along racial lines (Peterson
and Krivo 2010). Insofar as the mechanisms linking cohort characteristics to
criminal violence involve the capacity of such institutions to assimilate successive
generations, it follows that generations of Blacks and Whites have been raised
by families living in segregated neighborhoods, attended segregated schools, and
entered into segmented labor markets. The ability of such segregated institutions
to effectively socialize large cohorts of children is a defining factor in the
perpetuation of social inequalities, which have generated profound differences
in homicide rates across racial groups in the United States.
We would be remiss not to acknowledge several caveats with our analyses
that limit our ability to engage more directly with prior scholarship in this
area. For one, our empirical models necessitate age-by-race specific measures
of homicide and victimization. The SHR provides the longest-running source of
this information, but these data only extend back as far as 1975. While there
are clear advantages to using the SHR, our models do not directly align with
previous research (which has often relied on measures gleaned from the UCR
arrest records). Accordingly, the discrepancies between our findings for total
homicide offending and prior work may be due to broader issues with official
data collection over time. Because our analyses span different time frames,
we speculate that the differences in results for total homicide offending rates
may arise from differences in arrest reporting and subsequent disaggregation.
Regardless, our goal was not to dispute the pioneering work of others but rather
to explore more nuanced pathways through which race, population dynamics,
and segregation have influenced trends in criminal homicides in the United
States.
Relatedly, a growing body of literature demonstrates how exogenous shocks
and large-scale social changes can influence long-term crime trends (Rosenfeld
2018; Baumer, Vélez, and Rosenfeld 2018). When considered in the context
of the current study, we might anticipate such period effects to influence not
only homicide rates but to indirectly contribute to cohort characteristics by
differentially influencing the relative size and rates of nonmarital birth characterizing Black and White cohorts overtime. The most obvious examples are
the crack cocaine epidemic and the differential impact of mass incarceration on
communities of color. Indeed, both mass incarceration and the homicide boom of
the late 1980s had a disproportionate effect on specific cohorts of young, Black
men. While beyond the purview of the current study, it is entirely possible that
these period effects reshaped the long-term life chances of this cohort, far beyond
the historical moment which such events occurred.
We are also limited by our inability to examine the more proximate mechanisms linking cohort characteristics and criminal homicide, such as indicators
of school crowding, educational attainment, or labor market outcomes. And
indeed, the threats posed by omitted variable bias remain a considerable hurdle
in APCC models (O’Brien 2014). The most glaring omission in this regard is our
lack of measures of age-by-race specific unemployment rates, which would allow
us to examine directly whether labor market edging explains the strong effect of
cohort size on Black homicide rates. To our knowledge, no such measures exist
for the full period of observation used in our analyses. Rather than view this as
a critical limitation, we echo Robert K. Merton (1987) who argued that “before
one proceeds to explain or to interpret a phenomenon, it is advisable to establish
that the phenomenon actually exists, that it is enough of a regularity to require
and to allow explanation” (3). We view the empirical contributions described
here as a necessary first step toward confirming the differential impact of
population dynamics on violence within racial groups, thus laying the foundation
for further research into the mechanisms driving the disproportionate influence
of cohort characteristics on criminal homicide over the past 60 years.
Despite these caveats, our findings reaffirm the importance of systematically
incorporating demographic processes into criminological research. Social control
theories have long extolled the central role played by social institutions in
suppressing violent crime, but these theories have devoted little attention to the
ways in which rapid population growth might strain institutional capacity. Our
work further underscores the inexorable linkages between population dynamics
and institutional constraints in the propagation of racial inequality in the United
States. To paraphrase Richard Easterlin, year of birth indeed marks a generation
for life. In the context of criminal homicide in the United States, it is clear
that the enduring consequences of cohort characteristics for homicide offending
and victimization unfolded differently depending on race. Consistent with a
growing body of scholarship, the results presented here suggest that crime,
violence, and the perpetuation of racial inequality in the United States can
be best viewed through a historicized understanding of bifurcated population
dynamics.
Posts
No comments:
Post a Comment