Excerpted From: Ryan D. King and Michael T. Light, Have Racial and Ethnic Disparities in Sentencing Declined? , 48 Crime and Justice 365 (2019) (19 Footnotes and Reference List) (Full Document)
The past half century of American criminal justice is defined by two alarming trends. The first is the dramatic expansion of the size and scope of the criminal justice system. By any metric, be it police force size, prison construction, or dollars spent on the administration of justice, the system has ballooned since the 1970s. This expansion is most evident when we look at the number of persons in prisons and jails per capita. During the first three-quarters of the twentieth century, the imprisonment rate was consistently around 100 prisoners per 100,000 population, or closer to 150 when including both prison and jail inmates. Then, beginning in 1974, the total US incarceration rate began a striking ascent, increasing fivefold to 767 per 100,000 by 2007. This rate makes the United States an extreme outlier on the world stage, so much that a separate axis is needed when the American rate is graphed alongside those of western European and North American countries.
Along with this sea change came a second and equally troubling characteristic: vast racial disparities. At the front end of the justice system, blacks are arrested at far higher rates than whites (Heath 2014), and racial disparities in arrests for drug crimes increased markedly during the war on drugs of the late twentieth century (Mitchell and Caudy 2015, p. 93). At the back end, the proportion of African Americans in prison easily surpasses their proportion in the general population. According to one recent estimate, the imprisonment rate for African Americans is over five times that of whites, and in some states (e.g., Oklahoma) an incredible one in 15 black males over the age of 18 are in state prisons (The Sentencing Project 2016).
Much legal and social science scholarship attempts to explain why racial disparities exist in just about every criminal justice statistic and to determine what causes them. Three contributing factors are investigated: racial differences in offending, enactment of ostensibly race-neutral laws that have disparate racial impacts (e.g., the infamous federal crack vs. cocaine disparity; Tonry and Melewski 2008, pp. 29-30), and unequal treatment by the justice system. In other words, how much of the disparity is due to discrimination by those who legislate, arrest, prosecute, and punish?
This question is central to the subset of criminal justice research that focuses on sentencing. No issue has received more attention in the scholarly literature on sentencing than whether nonwhite defendants are treated more harshly than similarly situated whites. The answer is consequential in several ways. Most fundamentally, the Fourteenth Amendment guarantees equal protection under the law, and this core principle of fairness is called into question if race infects the decision to punish. Further, sending a convicted defendant to prison has ramifications well beyond the prison walls, as a bout of imprisonment adversely affects employment opportunities, wages, and long-term health (Pager 2003; Western 2006; Massoglia and Pridemore 2015). On that note, the sheer size of the imprisoned population and its racial disparities have motivated many scholars to claim that the criminal justice system is now a salient driver of racial stratification in the United States (Alexander 2012). As Wakefield and Uggen (2010, p. 401) state, “the prison has emerged as a powerful and often invisible institution that drives and shapes social inequality.” Differential treatment in the criminal justice system can also foment cynicism toward our legal and political authorities, which in turn is associated with higher crime rates (Tyler 1990; LaFree 1998; Roth 2009; Kirk and Papachristos 2011). It is thus critical to understand whether race matters in sentencing decisions, and if so, why?
Given gross disparities in arrest and aggregate imprisonment rates, it may seem axiomatic that blacks, and perhaps members of other racial minorities, are more likely than whites to be sentenced to prison. If the question is posed simply as Are whites and nonwhites sent to prison at different rates? the answer is unequivocally yes. The odds of imprisonment given a conviction in the federal courts are over three times higher for black than for white males, and the difference for Hispanic compared with non-Hispanic white males is even greater (Ulmer, Painter-Davis, and Tinik 2016). Comparable raw disparities in the likelihood of imprisonment and in the duration of prison sentences are apparent in Minnesota (King and Johnson 2016; Frase 2019), Delaware (MacDonald and Donnelly 2017), and Pennsylvania (Ulmer, Painter-Davis, and Tinik 2016). It is not controversial to state that sizable racial disparities in sentencing exist, akin to what we see in arrest and imprisonment statistics. The more complicated question, and one that generates less consensus, is whether unwarranted racial disparities in sentencing exist. That is, do racial differences in punishments merely reflect racial differences in crime severity and the criminal records of defendants?
It seemed for a time that this may be the case. Reviews of noncapital sentencing studies as of the early 1970s concluded that the relationship between race and sentencing outcomes was effectively nonexistent once proper controls for offense severity and criminal history were considered (Hagan 1973, p. 378; Kleck 1981). As Kleck wrote of studies published as of 1979, “only eight [of 40 studies] consistently support the racial discrimination hypothesis, while 12 are mixed and 20 produce evidence consistently contrary to the hypothesis” (p. 789). Kleck added that it was a distinct possibility that nonwhites (blacks in particular) were treated more leniently.
Fast-forward 25 years, and authors tasked with reviewing the state of sentencing research reach a different conclusion. A meta-analysis of 71 studies through the year 2002 concluded that, even after controlling for the seriousness of the criminal offense and the criminal history of the offender, blacks are, on average, sentenced more harshly than whites (Mitchell 2005). In other words, racial differences in sentencing outcomes are not explained by characteristics of the crime or the defendant's criminal record. Sandwiched in between the early summary reviews and later meta-analyses are several additional canvasses of the sentencing literature that reach various conclusions. Some found little evidence of unwarranted racial disparity (Pratt 1998). Others conclude that race continues to infect sentencing decisions and that the “'discrimination thesis' cannot be laid to rest” (Spohn 2000, p. 428; Sweeney and Haney 1992; Baumer 2013).
Our objective here is not to conduct another meta-analysis or exhaustive summary of research on racial disparities in sentencing. Our reading of the literature leads us to the same summary statement that Baumer (2013, p. 12) offered: “The consistent finding from recent overviews on studies of race and sentencing in samples of convicted defendants is that there are often relatively small but statistically significant direct race differences in the probability of imprisonment to the disadvantage of blacks (compared to whites), and comparatively smaller and statistically non-significant direct race differences in prison sentence lengths between these groups.”
Like Baumer, we note that racial disparity varies by jurisdiction, type of crime, and judge. Still, at this moment in history, the evidence suggests that racial disparities in sentencing exist and are not fully explained by the confluence of race with crime severity or the extensiveness of defendants' criminal records.
Rather than focus on the singular question of whether unwarranted disparities in sentencing exist in the aggregate, we give overdue attention to three larger themes. First, to what extent have sentencing disparities changed over time? Ironically, at the same time that studies in the 1970s found little evidence of unwarranted racial disparity in sentencing, sentencing reforms were being undertaken to eradicate racial disparities in sentencing by limiting discretion and increasing parity. Adding to the irony, studies of racial disparity after the implementation of sentencing guideline systems--for instance, in Pennsylvania, Minnesota, and the federal system--found evidence of racial disparities, although they appear to have diminished somewhat in Minnesota (Miethe and Moore 1985, p. 358). We cannot make perfect sense of this apparent disjuncture between research findings and policy shifts, but we can ask and try to answer a related question: have racial and ethnic disparities attenuated over time?
Second, it is time for a critical appraisal of theoretical explanations of race and sentencing. We identify four explanations that frequently motivate research on the topic, ranging from cognitive theories that emphasize implicit biases to macro-level theories that draw attention to county and state demographics. Not all theories square with the available evidence. For instance, the predictions of the oft-cited racial threat theory only sometimes pan out in empirical research. Other explanations, such as the focal concerns perspective on sentencing, arguably fit the data too well, raising questions about the theory's specificity and falsifiability. We weigh the merits of, and shortcomings with, theories of race and sentencing.
Third, we identify and expound on data problems that preclude a more complete assessment of the reasons for sentencing disparities. The off-the-shelf sentencing data used in most studies, including our own prior work, are inherently limited because of how they are collected. The data typically relate to convicted felony offenders who receive sentences, yet by that point in the process several earlier decisions have been made that almost certainly have consequences for sentencing outcomes. For instance, felony charges that result in misdemeanor convictions do not show up in felony sentencing files. Prosecutors' decisions to drop cases or to charge more severe crimes will affect which cases appear in the file and, if convicted, the ultimate sentence. Yet the exercise of prosecutorial discretion is not captured in most sentencing data. The few studies that follow cases from the point of arrest show that much of the racial disparity in sentencing reflects decisions made by prosecutors, particularly the decision whether to charge defendants under statutes that require mandatory minimum sentences (Rehavi and Starr 2014).
To accomplish these three objectives, we organize this essay into five sections.
Section I provides an overview of recent scholarship on race and sentencing. Most research continues to find that blacks are more likely than whites to receive prison sentences, even when accounting for racial differences in criminal history and crime severity. A similar pattern exists for Hispanics relative to non-Hispanic whites, a difference that is even more pronounced for Hispanic noncitizens.
In Section II, we turn to the question of whether racial and ethnic disparities in sentencing have changed during the past three decades. Our analysis of decades of data from the federal courts, Minnesota, and the State Court Processing Statistics Survey suggests some movement toward equality.
Section III appraises four frequently cited theories of racial disparity in sentencing: racial threat, focal concerns, implicit bias, and a loosely related set of ideas we combine under the heading “social distance.” Our goal is to evaluate each theory's logic and fit with the pattern of findings from prior research. In our assessment, the racial threat explanation is not well supported, and the focal concerns perspective lacks specificity. Tests of the focal concerns perspective tend to highlight confirmatory evidence while minimizing findings that are inconsistent with predictions.
Section IV discusses the limitations of what Baumer (2013) calls the modal approach to sentencing, that is, the study of racial disparities based on samples of convicted felons provided by sentencing commissions. Gaining clearer understanding will require new data. The lack of data that follow cases from the point of arrest through to their final dispositions, and the near absence of victim information in noncapital sentencing research, impede the ability to explain why racial disparities exist.
Section V concludes with recommendations for future research. Two research issues warrant further attention. First, more work is needed on trends in disparities. We were surprised at how little has been written about changes over time. Our analysis in Section II provides a descriptive account but does not fully answer the question of why sentencing disparities appear to be declining. To this end, future work should take a hard look at policy changes that affect the presumptive sentence and use of criminal histories.
[. . .]
The findings reported and conclusions drawn in this essay have implications for future research and sentencing policy. We emphasize four.
First, our finding that racial and ethnic disparity is declining warrants further scrutiny and analysis. For instance, do the trends hold up in states other than those we looked at? And do they persist when different methodological choices are made? We can imagine different methods to calculate adjusted disparity, different ways to measure the presumptive sentence, and more sophisticated modeling of change over time that takes other case characteristics into account. Replications are needed before offering answers to a related important question: if disparities have attenuated, what explains the change?
Our descriptive analyses of federal and Minnesota data offer some clues, but much work lies ahead to pinpoint the reasons. Our analysis, like Frase's (2009) analysis of racial disparities in prison, indicates that racial differences in presumptive sentences cause much of the disparity. Figures 2, 4, 9, and 10 show that trends in pronounced sentences tether closely to trends in presumptive sentences. The adjusted disparity analyses suggest some additional reduction in racial and ethnic disparities, but in guideline systems the recommended sentence drives much of the change.
Second, our analyses do not estimate the precise contributions of the various factors that feed into presumptive sentences, such as sentencing policy, conviction offenses, and criminal histories. Future research could add specificity by decomposing the relative contributions of each, particularly as they contribute to changes in disparities over time.
Third, theories in use need elaboration and augmentation. Racial and ethnic disparities vary across jurisdictions and over time. A cogent explanation of racial disparity should incorporate these observations. We doubt that many of the most-often-cited theories accomplish this, although arguments stressing diversity of the workforce have promise for explaining disparities that are not attributable to racial differences in presumptive sentences, criminal histories, or crime severity (Ward, Farrell, and Rousseau 2009). Research on whether a diverse courtroom workgroup mitigates racial and ethnic sentencing disparities warrants high priority.
Finally, we believe that factors that contribute to determination of presumptive sentences in guidelines systems warrant further examination. Disparities in federal sentencing, figure 16 shows, appear to decline following changes in policies that have racially disparate impacts. Other factors such as criminal history (see Frase 2009, p. 265) that vary by race and partly determine presumptive sentences should also be investigated. If we better understood which factors most contribute to racial differences in presumptive sentences, we could better understand the causes of inequities and whether apparent declines in racial disparity are likely to persist.
Ryan D. King is a professor of sociology at The Ohio State University.
Michael T. Light is an associate professor of sociology and Chicano/Latino studies at the University of Wisconsin--Madison.