Abstract
How do members of racial and ethnic groups explain the origins of unpaid legal debt from monetary sanctions, and how do such attributions undergird group differences in support for policy responses that escalate punishment? Using data from the Chicago Area Finances Survey, 2019, we apply an attributional typology of stratification beliefs to account for why legal debt from fines, fees, and tickets goes unpaid. We find differences in attribution types along key measures of socio-demographics and political values, and we identify racial differences in these attributions when other measures are held constant. How people understand why legal debt goes unpaid shapes their policy preferences as well, and they explain a small but significant fraction of racial and ethnic differences in the desire for punitive recourse.
Keywords
Following the fatal shooting of Michael Brown in 2014, and the subsequent unrest in Ferguson, Missouri, the U.S. Department of Justice (DOJ) initiated an investigation to understand how a local police department could be at odds with the community’s majority-Black population whom officers had sworn to protect. The findings uncovered that the police department, in collusion with the city government and municipal court system, prioritized revenue streams from fines, fees, and tickets above concerns of public safety (U.S. DOJ, Civil Rights Division 2015). As much as one fifth of the 2013 local budget derived from these sanctions, for example, making it the second-largest source of local revenue (ahead of property taxes but behind sales taxes). These revenues nearly tripled from $1 million to $2.6 million over 10 years (2003–2013), even as the city’s population decreased by more than 10 percent during the same period. Moreover, although Black residents comprised two thirds of Ferguson’s population, they accounted for 90 percent of all citations. If headlines are any indication, the DOJ findings shocked many across the country as a damning revelation (Page and Soss 2018). Residents of Ferguson and people of color across America, however, likely viewed the findings as confirmation of experiences they know well.
The exploitive fines, fees, and ticketing practices that were brought to light by “Ferguson” can elicit polarized reactions from different ethnoracial groups. These differences result, in part, from unique encounters with the justice system (Bobo and Johnson 2004; Peffley, Hurwitz, and Mondak 2017) as well as distinct takes on why some people have difficulty paying monetary sanctions. Those who believe that unpaid legal debt is a consequence of bad judgment, poor money management, or ruthless defiance of the law could be inclined to support policy responses that escalate punishment. By contrast, those who see this debt as a consequence of wider social problems, like underemployment, unlivable wages, or targeted surveillance, may shudder responses of reprimand begetting reprimand. If distinctive ways of thinking about the origins of unpaid legal debt vary by race and ethnicity, then groups may be worlds apart in the basic meanings they attach to unpaid monetary sanctions and how the state should respond. In the analysis to come, we shed light on this possibility using data from the Chicago Area Finances Survey (CAFS), 2019, which allows us to measure the extent, sources, and political consequences of differences among Blacks, Latinxs, and whites in their attributions behind the sources of unpaid fines, fees, and tickets.
Explaining Unpaid Legal Debt under the Shadow Carceral State
Stratification beliefs can be defined as the study of “what people believe about who gets what and why” (Kluegel and Smith 1981:30). In answering the “why” question, people tend to stress dispositional (internal) or systemic (external) attributions in explaining the sources of social problems like poverty (Feagin 1972, 1975), criminal behavior (Thompson and Bobo 2011), and so on. Whereas dispositional explanations locate blame within the poor themselves (e.g., lack of work ethic, moral deficiencies), a systemic lens sees poverty as a consequence of the economy in which the poor live (e.g., underemployment). Decades ago, James R. Kluegel and Eliot R. Smith (1981) lamented the absence of people of color in research on stratification beliefs. Although some studies are now available to correct this absence (e.g., Hunt 2007; Kluegel 1990), work that delineates the complicated intersections between race and ethnicity, stratification beliefs, and their political consequences remains in its infancy (Hunt 2016). Included in this category is the topic of public responses to justice-related policy (Mullinix and Norris 2019; Thompson and Bobo 2011). Even fewer studies attempt to uncover how stratification beliefs explain the origins of unpaid legal debt and the ways these attributions inform support for policy responses.
Why are monetary sanctions a vital topic to consider in the current historical moment? For starters, the number of Americans they affect far surpasses the incarcerated or convicted population (Martin et al. 2018; Page and Soss 2018). As many as 10 million Americans, who are disproportionately poor and people of color, owe more than $50 billion in debt from fines, fees, and tickets (Harris, Evans, and Beckett 2010; see also Harris 2016), and revenues from these sanctions have increased nearly 650 percent since the late 1970s despite crime trends being at 30-year lows (Henricks and Harvey 2017). Beyond their growth, these sanctions are redefining the scope of U.S. punishment. They supplement prison-based sentences with an intensification of alternative punishments that can be imposed upon people outside of penitentiary walls (Beckett and Murakawa 2012). Their money-generating capacities go further to reconfigure the justice system by migrating into it administrative tasks that are associated with taxation and debt collection. These financial penalties blend banal features of bureaucracy (e.g., multijurisdictional divisions of labor, scripted exchanges between those who owe and collect legal debt, constant monitor of payment) with routine extraction backed by the threat of penal coercion.
Worlds of Difference in Explaining Unpaid Legal Debt
Where people stand on justice-related issues is often shaped by their social position. Ethnoracial groups live in “different realities” when it comes to how they see the justice system (Bobo and Johnson 2004; Peffley et al. 2017). Because whites tend to evaluate the justice system as “colorblind,” they are unlikely to attribute unequal outcomes created by that system as a result of bias or discrimination (Bonilla-Silva 2003). Frameworks referred to as minority threat, group position, and conflict theory make sense of these views by stressing rationalized interests that justify either how the world is or ought to be (Bobo 1999; Fox 2004). Those atop the ethnoracial order see those at the bottom as competitive threats for valued resources, and they defend their advantage by scapegoating minoritized groups for their own status. Inequality-making practices, from stop-and-frisk to mass imprisonment, become justifiable strategies under a supposed impartial system, ones that rationalize policing a range of unlawful deeds committed by those undeserving of public compassion. Knowing that policy preferences tend to be group-centric, influenced “in powerful ways by the attitudes citizens possess toward the social groups they see as the principal beneficiaries (or victims) of the policy” (Nelson and Kinder 1996:1055–1056), we seek to extend this work into public opinion on punishment involving legal debt.
Why can some people satisfy their debt from monetary sanctions while others cannot? To what extent do dispositional or systemic attributions influence these answers? Are there ethnoracial differences in these explanations? The first portion of our analyses answers these questions. Prior attributional research leads us to anticipate Black, Latinx, and white Chicagoans to be worlds apart in their explanations for why legal debt remains unpaid. Why? In the Chicago area especially, race and ethnicity continue to color most every available metric of inequality, where whites sit atop the stratified order and Blacks on the bottom with Latinxs situated in-between. Group position not only structures differential interactions with major institutions like education, housing, and the justice system, but it likely shapes how Blacks, Latinxs, and whites draw fundamentally unique conclusions about outstanding debt from monetary sanctions.
Alternative Explanations for Why People Pay Legal Debts (or Don’t)
The second portion of our analyses explores whether ethnoracial differences in attributions of unpaid legal debt can be explained by other factors. Several alternatives may account for how people understand unpaid legal debt and their preferred policy responses. Studies show, for example, that political values tilting more conservative correspond with “get tough” stances on punishment, including policies like capital punishment (Jacobs and Carmichael 2002; Soss, Langbein, and Metelko 2003). The idea is that an emphasis on reprimand is consistent with conservative values of accountability. “Law-breakers” are seen as rational individuals responsible for their own choices, deserving of whatever consequences result. These political views borrow “free enterprise” concepts from a market economy, where “[p]unishment should be equivalent to the offense, so that justice consists in a kind of equity or fair trading which exchanges one harmful act for another which equals it” (Garland 1990:113).
If we take insights about political values and apply them to the escalation of punishment for unpaid legal debt, it becomes reasonable to suspect that the mantra of personal responsibility is among the factors linking conservativism with punitiveness more generally. The hypothesis can be formalized accordingly:
Aside from political values, another known factor shaping how people understand justice-related matters is exposure to crime. Mary Pattillo (2007) points out, “People live locally. The most earnest political battles are played out when they face threats to their neighborhoods or try to fashion a new kind of neighborhood” (p. 2). Residents understand these threats to have wide-ranging consequences, ranging from driving down their home values to feeling unsafe. Of course, stereotypes color these perceptions of crime exposure. Both Blacks and whites are more likely to view neighborhoods as dangerous, high crime or not, when young Black men reside in the area (Quillian and Pager 2001). Whites tend to overestimate the degree that Black people are involved in street crime and exaggerate their own risk of victimization (Pickett et al. 2012), but even residents in majority-Black spaces have supported “tough on crime” stances out of desires for safety that later imprisoned many from their own communities (Forman 2017).
Exposure to crime may predispose individuals to understand the sources of crime-related behavior in different ways. Those who reside (or think they reside) in high-crime areas can be motivated to adopt dispositional attributions for why people commit unlawful acts, and by extension, deservedly face punishments like monetary sanctions (Thompson and Bobo 2011). Debt tied to criminal behavior becomes seen as a justifiable consequence under these circumstances. It becomes moral-turned-financial atonement for the legal sinner in need of repentance. The hypothesis is as follows:
Do Stratification Beliefs Mediate Ethnoracial Differences in Policy Responses?
The third portion of our analyses seeks to understand if stratification beliefs are of political consequence. Do they shape the public’s desire to escalate punishment for those with unpaid legal debt? Might they also mediate ethnoracial differences in support for these policies? Previous studies show that dispositional accounts of poverty predict lower support levels for welfare spending, while systemic orientations increase support (Feagin 1975; Kluegel and Smith 1986). Other studies find consistent patterns for diversity initiatives designed to ameliorate inequalities in workplace representation (Scarborough, Lambouths, and Holbrook 2019). Structural outlooks are associated with support for redistributive and compensatory policies, while individualistic explanations are associated with their opposition. When it comes to justice-related policies—including the death penalty, increased spending on police and prisons, and adopting more lenient parole boards—recent work confirms similar patterns (Peffley et al. 2017; Thompson and Bobo 2011). Those who identify systemic factors as the causes of crime are more likely to support liberal policies of punishment, whereas those who identify with dispositional accounts are more likely to sing a punitive tenor.
While policy preferences tend to be group-centric (Nelson and Kinder 1996), it is worth noting that a systemic-dispositional divide does not discretely map onto ethnoracial lines. Blacks are more likely than whites to identify with structural interpretations of poverty, but Blacks and whites adhere to dispositional orientations at similar rates (Feagin 1972; Kluegel and Smith 1986). These patterns may reflect the taken-for-granted hegemony of individualism, an ideology that yields “common sense” thinking for most Americans. Rather than see individualism as an alternative to structuralism, researchers stress the two can be layered onto one another as a dual consciousness. The simultaneous identification with individualism and structuralism is especially prevalent among minority groups like Blacks and Latinxs (Hunt 1996, 2007; Sue and Lambert 2020).Those who have been historically discriminated against are less convinced by dominant ideology and more likely to articulate system-challenging alternatives, but within-group fragmentation among minoritized groups tends to blur the cohesion of counterviews (Mann 1970).
Given the attributional variation that can occur within and across groups, we are left asking two competing questions about how we might understand ethnoracial differences in policy preferences that further punish the indebted. Are these differences fundamentally a product of where individuals and their groups are located within the ethnoracial order? Or might intragroup variation in attributions of unpaid legal debt elucidate differences between groups? We will answer both questions. If attributional accounts among Blacks, Latinxs, and whites are consistent across subgroups of these populations, and these attributions correspond with particular policy preferences, then common beliefs about the sources of unpaid legal debt will help to explain some of these groups’ polarized views. In other words, we anticipate stratification beliefs about unpaid legal debt to mediate ethnoracial differences on punitive responses directed at those who either cannot or do not pay the monetary sanctions the state claims they owe.
The CAFS, 2019
Do stratification beliefs play a mediating role between ethnoracial location and policy views on unpaid legal debt? If so, to what extent do these different explanatory modes act as an intervening variable on policy responses involving monetary sanctions? We rely upon the CAFS, 2019, to answer these questions. The CAFS project is a study designed by the primary author to understand how residents of the Chicago area feel about a variety of issues related to local government finance. Not only does it feature original survey items on the beliefs behind why debt from monetary sanctions goes unpaid, but it also contains measures on people’s sociopolitical ideologies, their sense and fear of local crime, and general views on punishment.
Although many features make the Chicago area unique from other cities across America, there are at least three reasons why it is well-suited to fulfill our objectives. One, the Chicago area, and Illinois generally, is center focus of a growing body of research on monetary sanctions. This work has been completed by think tank analysts (Chicago Appleseed and Chicago Council of Lawyers 2020; The Woodstock Institute 2018), journalists (Ramos 2019; Sanchez 2018; Sanchez and Ramos 2018), and academics (Friedman 2021; Friedman and Pattillo 2019; Pacewicz and Robinson 2020) alike, and it allows our findings to be contextualized more broadly. Two, our fieldsite has been an area for reform on fines and fees at the state level and parking tickets at the local level (e.g., the Criminal and Traffic Assessment Act of 2019). And three, finance policies taking root in Chicago, from public-private partnerships to the monetization of government services, have a history of preceding policy change later diffused across the country.
Survey participants were selected from an online panel maintained by Qualtrics. While the general adult population in the sampling frame was invited to complete the questionnaire, a screener set of questions was used to decide eligibility. Respondents were eligible if they resided in Cook County, Illinois, the county seat of Chicago, and quotas were set on measures of race and ethnicity, income, and education. These inclusion criteria were purposefully chosen because many of the survey items probe at matters of race, ethnicity, and socioeconomic status. Although the CAFS project is an opt-in survey, meaning that it follows a non-probability sampling strategy, targets were matched with census data to ensure characteristics of the obtained sample reflect the underlying population.
A sample size of 1,203 was obtained. Because 9.48 percent of these cases included data that are missing, we use a listwise exclusion method to retain only those cases with complete observations (n = 1,089). The ethnoracial profile in our final analysis is as follows: 30.21 percent Black, 36.46 percent Latinx, and 33.33 percent white. While this sampling process is less sophisticated than a true probabilistic design, we are confident the heterogeneity of participants recruited through this strategy is sufficient enough to fulfill our exploratory research aims.
Lay Accounts for Outstanding Debt from Monetary Sanctions
To ascertain accounts for why debt from monetary sanctions goes unpaid, the CAFS project posed four questions to respondents, in an agree/disagree format, specifying potential root causes along a dispositional-systemic continuum. These four statements, along with their response distributions, are shown in Table 1. Respondents were asked to what extent they agree with the following individualist attributions:
People do not pay their legal debt when convicted of a crime because they would rather not work (e.g., laziness).
Debt remains unpaid because people do not care.
They were also asked to what extent they agree with these structuralist attributions:
Lack of decent employment causes debt to remain unpaid.
Poverty causes unpaid debt.
Reasons People Cannot or Do Not Pay Their Legal Debts: Individual versus Structural Attributions by Race and Ethnicity.
Generally speaking, Blacks are least likely among the three groups to adhere to individualism. Whites are most likely to identify with individualism and least likely to identify with structuralism. Consistent with other findings (Sue and Lambert 2020), Latinxs adhere to mixed views. Their perspectives are consistent with Blacks on questions of structuralism. For questions of individualism, their response patterns mirror whites on one item and fall in-between Blacks and whites on another.
The least widely endorsed attributions among survey respondents were individualist ones. This finding departs from national-level studies that confirm individualism as the dominant paradigm for understanding social problems (Feagin 1972, 1975; Kluegel and Smith 1986). The item with the least amount of support is, “people do not pay their legal debt when convicted of a crime because they would rather not work (e.g., laziness).” No majority of any group supported this individualist stance, including 28.57 percent of Blacks, 46.35 percent of Latinxs, and 47.11 percent of whites. The other individualist item, “debt remains unpaid because people do not care,” prompted distinct responses among all three groups. While over a third of Blacks agreed with this statement, more than half of Latinxs and two thirds of whites offered their support.
Meanwhile, the most widely endorsed attributions were structural ones. The item with the most support includes “poverty causes unpaid debt.” Nearly two thirds of Blacks and Latinxs and about half of whites adhere to this view. The second most frequently claimed attribution is “lack of decent employment causes debt to remain unpaid.” Two thirds of both Blacks and Latinxs and two fifths of whites (38.57 percent) endorse this view. On both these structural items, double-digit spreads separate respondents of color from their white counterparts.
Those who participated in the CAFS project adhere to any number of explanations for why debt from monetary sanctions remains unfulfilled, and many of their responses are neither consistent nor mutually exclusive. Significant portions of all three groups simultaneously hold onto dispositional and systemic attributions. The majority of Latinxs identify with structuralist items as well as one individualist item at the same time. Their views reflect a certain level of overlap that accommodates many outlooks. This is to say that explanations for unpaid legal debt are multifaceted, if not contradictory, in ways that lend themselves to no single orientation. To discern where people fall along the individualist-structuralist continuum, we follow the lead of James R. Kluegel (1990) to classify a joint configuration of these responses. Whereas other researchers have applied such a classification scheme to a broad array of topics, including the sources of poverty and criminal behavior, our own categorization scheme is unique. It takes up the subject of how people explain why people cannot or do not make good on sanctions they owe the state.
Attributional Modes of Explaining Outstanding Debt
The attributional approach to mapping stratification beliefs follows a “compound belief” strategy of measurement. Instead of attempting to sort individuals along an individual/structural binary, the strategy embraces a measurement scheme that blends individualist and structural endorsements or evades them altogether. Given that minorities are more likely to embrace interpretations that fuse together individualism and structuralism (Hunt 1996, 2007), this approach may be especially useful for capturing a more nuanced take on how people understand inequality. It includes a typology of response patterns across multiple survey items that features four categories: (1) those who consistently identify with individualism, (2) those who consistently identify with structuralism, (3) those who adhere to mixed views of individualism and structuralism, and (4) those who reject orientations of individualism and structuralism altogether.
Our attribution categories are defined as follows. Respondents are classified as individualists or structuralists if their responses correspond to either of these modes in a consistent direction. Individualists include those who either (1) endorse neither of the structuralist attributions and at least one of the individualist attributions or (2) endorse both individualist items but also one of the structuralist attributions. Structuralists, by contrast, are labeled as such if they either (1) endorse neither of the individualist attributions and at least one of the structuralist attributions or (2) endorse both structuralist items but also one of the individualist attributions. We reserve the “mixed” category for respondents who balance individualism and structuralism. Respondents are classified as mixed if they either (1) affirmed one of the individualist and structuralist items but denied the other categories or (2) affirmed all the items for both individualism and structuralism. Those respondents labeled under the category “none” include those who reject individualist and structuralist attributions in consistent fashion.
How Do Different Ethnoracial Groups Understand the Origins of Unpaid Legal Debt?
Table 2 presents the attributions for outstanding legal debt by ethnoracial group. It confirms some of the preliminary trends observed in Table 1. Less than a third of our respondents adhere to individualist interpretations. Among those who hold dispositional stances, whites comprise over half of them. How does white adherence to individualism compare with other groups? Whites are more prone to identify with individualism, albeit not a majority of the group, than are Blacks and Latinxs. In fact, they identify with individualism over structuralism by a factor of 1.55, whereas Blacks and Latinxs identify with structuralism over individualism by factors of 3.56 and 1.93, respectively. When we combine individualists with mixed-mode responses, which includes partial adherence to individualism, whites are the only group to surpass the majority threshold (57.81 percent) on these two explanatory modes. That said, sizable portions of both Blacks and Latinxs support dispositional takes on unpaid legal debt or a blend of individualist and structuralist stances (Latinxs = 38.13 percent; Blacks = 26.69 percent).
Response Patterns for Explanatory Modes of Unpaid Legal Debt by Race and Ethnicity.
The most frequently observed attribution, although still not a majority, includes those who endorse a structural framework. Nearly half of our sample identify with systemic reasons for why people either cannot or do not satisfy their legal debts. Among those who adhered to these explanations, Blacks and Latinxs are the largest groups to comprise this category. Whites are least likely to adhere to structuralism. Of the remaining attributions, mixed and none, distinctions fade. Approximately 10 percent of each group is classified as either adhering to mixed views or rejecting individualism and structuralism on the whole. What these trends indicate is that group differences are more apparent among the individualist and structuralist orientations than with the categories “mixed” and “none.” These patterns are of significance in that the explanatory mode of stratification beliefs may be a key mediator in explaining why Blacks, Latinxs, and whites assume different positions on escalating punishment for those with unpaid legal debt.
Factors Affecting Explanations among Blacks, Latinxs, and Whites
Table 3 shows descriptive statistics for each explanatory mode, in addition to variables researchers commonly associate with stratification beliefs. These covariates include measures of demographics, values and ideology, and crime salience factors. See the appendix for how these variables are operationalized. Are there ethnoracial differences on these items? Answering this question can shed light on whether ethnoracial differences toward policy preferences on monetary sanctions is mediated by mode of explanation or if these stances are confounded by other factors. The statistical tests of mean differences confirm what was suggested in Table 2. Ethnoracial differences are significant on two explanatory modes, individualism and structuralism.
Measures of Means, Standard Deviations, and Group Differences by Race and Ethnicity.
Black and Latinx means are significantly different.
Black and white means are significantly different.
Latinx and white means are significantly different.
*p < .05. **p < .01. ***p < .001 (two-tailed tests).
Statistical tests confirm significant differences among Blacks, Latinxs, and whites on measures of demographics, values and ideology, and crime salience. Group differences are significant for every demographic variable considered: gender, age, education, and income. For indicators of values and ideology, differences are significant for respondents who identify as Democratic Socialist, Democrat, Independent, and Republican. They are also significant on our measure of conservatism. When we consider crime salience, every measure registers a significant result. Group differences on a person’s sense of crime include an average of two measures, one that accounts for perceptions of neighborhood crime and another that considers perceptions of whether their neighborhood is safe to raise children. Also significant are differences in fear of crime, measured by whether respondents feel unsafe in their neighborhood. Last, there are significant differences on whether the respondent, or an immediate family member, has ever been a victim of crime.
Group differences point to the unique experiences among Blacks, Latinxs, and whites in terms of their backgrounds, value systems, and crime salience. The average Black respondent, for example, is considerably older than the average Latinx respondent but younger than the average white respondent. Moreover, the average Black respondent has less education attainment and earns less income than the average Latinx and white respondents but is more likely to identify as a Democrat and have been a victim of crime. Why do these correlates matter? They matter because these factors tend to be associated with particular stances on punishment. Thus, ethnoracial gaps in social identities, political orientation, and exposure to crime may have ramifications for how different groups come to understand why debt from monetary sanctions goes unpaid and what government ought to do about it.
Differences in Stratification Beliefs about Unpaid Legal Debt by Race and Ethnicity
To what extent do stratification beliefs mediate ethnoracial differences on policies that escalate punishment for unpaid legal debt? Our strategy for testing this relationship begins with determining the likelihood that attributional modes differ by ethnoracial group, holding constant indicators of social background, political values, and crime salience factors. Table 4 presents results from our multinomial logistic regression models. The dependent variable for these models is treated as a categorical variable indicating each explanatory mode using structuralism as the reference. That is, we look at how attributional modes differ by ethnoracial group compared with the structuralist mode of explanation. This approach allows us to understand how ethnoracial groups differ in their explanations of outstanding debt while accounting for the independence of unique attributional modes. Results described below are listed in relative risk ratios (RRR), which indicate how the risk of the outcome falling in the comparison group relative to the risk of falling in the referent group increases as the variable increases.
Multinomial Logistic Regression Models for Each Mode of Explanation.
Note. The main entries are relative risk ratios.
The reference category includes whites.
The reference category includes those who identify with Democratic Socialists, the Green Party, and the Democratic Party.
*p < .05. **p < .01. ***p < .001 (two-tailed tests).
At least four patterns stand out from our models that pool all ethnoracial groups. First, differences among Blacks and whites are significant in our models that predict no attributions and individualist attributions of unpaid legal debt. Compared with whites, Blacks are less likely to report no mode of explanation (RRR: 0.47, p < .01) and less likely to be individualists (RRR: 0.31, p < .001). Second, the only other demographic items that deliver salient results are age and education. Older respondents are more likely to adhere to individualism (RRR: 1.03, p < .001), while respondents with higher levels of education are less likely to adhere to individualism (RRR: 0.82, p < .01). Third, political values and ideology matter. Compared with those who identify with left or liberal parties, Republicans are more likely to cite dispositional factors for unpaid debt (RRR: 2.71, p < .01). Also, higher levels of conservatism return significant results on all modes of explanation. The more conservative respondents are, the more likely they are to adhere to no explanation (RRR: 1.53, p < .001), individualism (RRR: 1.74, p < .001), or mixed views (RRR: 1.64, p < .001). Fourth, respondents who have been crime victims or know a crime victim are less likely to reject all the explanatory modes (RRR: 0.44, p < .001).
While results from the pooled models confirm all groups provide different attributions for unpaid legal debt, particularly between Blacks and whites on individualism and structuralism, one question remains: Are determinants within each of these groups distinctive in predicting these attributions? We answer this question by narrowing our analysis to each group. Looking to our sociodemographic measures first, age predicts individualism among Blacks (RRR: 1.03, p < 05), Latinxs (RRR: 1.05,p < .001), and whites (RRR: 1.02, p < .05). It also predicts mixed views among whites (RRR: 1.05,p < .01). Education has more influence on whites’ stratification beliefs, and to a lesser extent Latinxs’, than it does beliefs held by Black respondents. The risk ratio of adhering to no explanatory mode decreases by 26 percent among whites (RRR: 0.74, p < .05) for every increase in education level, while the ratio of adhering to individualism decline by 27 percent among whites (RRR: 0.73, p < .01) and 32 percent among Latinxs (RRR: 0.68, p < .01) for increases in education. And income is a significant predictor of individualism among Latinxs, where higher incomes correspond with greater levels of adherence (RRR: 1.19, p < .05).
When it comes to political orientation, conservatism predicts two out of three attributions among Blacks and is associated with multiple explanatory modes for Latinxs and whites. That is, higher levels of conservatism among Blacks correspond to a higher likelihood of adhering to no explanation (None RRR: 1.35, p < .05; Individualist RRR: 1.47, p < .05). For Latinxs, higher levels of conservatism are associated with higher likelihood of reporting no mode (RRR: 1.43, p < .01), individualism (RRR: 1.48, p < .01), and mixed views (RRR: 2.20, p < .001). For whites, higher levels of conservatism correspond to higher likelihood of reporting no mode (RRR: 2.15, p < .05), individualism (RRR: 2.56, p < .001), and mixed views (RRR:1.84, p < .01).
Last, two measures of the crime salience factors are influential, albeit in inconsistent ways. Members from all three groups who have been crime victims or know crime victims are less likely to reject all explanatory modes. More specifically, Black respondents are 48 percent less likely to adhere to no attribution (RRR: 0.52, p < .05). Latinxs are 57 percent less likely (RRR: 0.43, p < .01), and whites are 61 percent less likely (RRR: 0.39, p < .01). When whites sense crime is high in their neighborhood, they are more likely to adhere mixed views than those who felt crime is a non-issue (RRR: 2.35, p < .05). In the end, what do these results from Table 4 tell us? They tell us that discrepancies among Blacks, Latinxs, and whites exist for how they understand the origins of unpaid legal debt.
Should Punishment Beget Punishment?
Differences in stratification beliefs may shape distinctive levels of support among Blacks, Latinxs, and whites when it comes to policy responses for unpaid legal debt. To determine the extent this statement is true, we explore two sets of questions that ask respondents whether outstanding monetary sanctions should beget other forms of exclusion or punishment. The first set of questions begins with the following prompt: Thinking about people who owe legal debts for breaking the law, some have suggested these debtors face additional penalties until the money they owe to the government is paid. Please tell me whether you believe the following rights or privileges should be restricted for those who have outstanding legal debt.
Respondents who agree were instructed to indicate a “yes” response on these items:
Eligibility for housing assistance
Eligibility for nutritional supplements (“food stamps”)
Eligibility for medical care assistance (“Medicaid”)
Ability to obtain financial aid for education
The second set of questions begins with the following stem: Are there any situations in which you would approve of the government confiscating the resources of those who have outstanding legal court debt?
If respondents agree that certain situations warrant government confiscation, they were instructed to indicate a “yes” response on these items:
If the person regularly missed debt payments?
If the person received some form of government assistance?
If the person had the money for debt payments but spent it elsewhere?
If the person had been allegedly involved in other illegal activity?
Both sets of questions probe at policies that possibly multiply the in-roads to punishment and entrench a wider net of surveillance and social control (Beckett and Murakawa 2012).
Are there differences among Blacks, Latinxs, and whites when it comes to punitive recourse on those with unpaid legal debt? Table 5 answers this question in the affirmative. It shows the rate of support by ethnoracial group as well as provides rate-ratio comparisons between groups. A ratio equivalent to the value of “1” indicates parity among the two groups compared. Values below “1” can be interpreted as lower rates of support than whites, while values above “1” can be interpreted as higher rates of support than whites. The findings show no parity on any single item. Black respondents are the least likely among the three groups to support punitive recourse for unpaid legal debt. Only one in three Black respondents (or less) support limiting the social safety net for debtors (min. = 27.66 percent, max. = 33.74 percent). Latinxs and whites, on the contrary, alternate between voicing the highest rates of retaliatory support. Latinxs score the highest values, although still a minority in absolute terms, on three of the eight items (Q1 = 39.29 percent; Q2 = 39.29 percent; Q3 = 37.78 percent), while whites score the highest values on the remaining five items (Q4 = 44.08 percent; Q5 = 58.40 percent; Q6 = 46.83 percent; Q7 = 82.09 percent; Q8 = 84.57 percent).
Should Punishment Beget Punishment? Possible Recourse for Outstanding Legal Debt in the Chicago Area by Race and Ethnicity.
It is worth noting that the comparison gaps presented in Table 5 range broadly between single- and double-digit spreads (min. difference = 1.01, max. difference = 17.95) and are more substantial on particular policy responses. Respondents are far more supportive of government expropriation than they are of policies that shrink the social safety net. Whereas no survey item relating to social safety net restrictions clears majority support from any ethnoracial group, these groups express majority support on multiple items relating to government confiscation. A majority of both Latinxs and whites express their support on three out of four of these items, and a majority of Blacks express support on two out of the four. The broadest approval gap exists between Blacks and whites (Black-white ratio = 0.62) on whether government should confiscate resources from those with outstanding debts if that person receives public assistance (Blacks = 28.88 percent approve, whites = 46.83 percent approve). Meanwhile, the narrowest gap exists between Latinxs and whites (Latinx-white ratio = 0.98) on whether access to financial aid for education should be restricted for those with outstanding legal debt (Latinxs = 43.07 percent approve, whites = 44.08 percent approve). While it is possible to identify multiple sources for these differences, one plausible explanation for these support gaps is that Blacks, Latinxs, and whites adhere to fundamentally different stratification beliefs for why debt goes delinquent to begin with.
Shrinking the Social Safety Net for Debtors
For the first set of questions that ask whether the social safety net should be withdrawn from those with outstanding legal debt, we use binary logit models to predict support (1 = yes, 0 = no) for stripping access to the following services: housing assistance, nutritional supplements (i.e., “food stamps”), medical care (“Medicaid”), and financial aid for education. These questions come at a moment when the welfare state is being eroded as a public institution, including but not limited to cuts to social programs and public sector jobs, a reduction in infrastructure upkeep, the sale of public resources to private interests, and the weakening of regulatory power in general. Each of our dependent variables includes two models. The first is a base model that discerns different support levels among ethnoracial groups, while factors of demographics, values and ideology, and crime salience are held constant. The second model builds upon the first by adding stratification beliefs as mediators. We evaluate these models and compare their efficiency using the Akaike information criterion (AIC) to indicate predictive validity based on our models’ deviance levels. Unlike measures of pseudo R2, the AIC is a model fit statistic that penalizes models for the inclusion of additional parameters. A reduction of at least 2.5 is generally considered the minimum threshold for a better fitting candidate model, while a reduction of 10 or larger indicates a clear model preference.
The two-step modeling process allows us to confirm (1) whether ethnoracial differences predict support for the escalation of punishment and (2) if explanatory modes act as intervening variables that explain away polarization between groups. Table 6 presents the results, and two trends merit emphasis. First, differences between Blacks and whites are statistically significant, net all other variables, in every base model. Blacks are less likely than whites to support harsh policies that further punish those with unpaid legal debt. Second, Black-white differences are reduced, although still salient on three of the four outcomes, once stratification beliefs are considered. These findings suggest that stratification beliefs, particularly individualism, affect preferences toward punitive recourse and that these accounts explain a small but significant portion of Black-white differences on harsh responses to unpaid legal debt. Results described below are listed in odds ratios (OR), which are unique from RRRs in that they are comparing the ratio of two odds as opposed to two probabilities.
Do Modes of Explanation Mediate Ethnoracial Differences in Attitudes toward the Escalation of Punishment for Unpaid Legal Debt?
Note. The main entries are odds ratios. AIC = Akaike information criterion.
The reference category includes whites.
The reference category includes those who identify with the Democratic Party, Democratic Socialists, and the Green Party.
The reference category includes those classified as structuralists.
*p < .05. **p < .01. ***p < .001 (two-tailed tests).
Models 1a and 1b predict support for restricting access to government housing assistance. Model 1a shows that Blacks are less likely than whites, by a factor of 0.62 (p < .05), to support this policy. The only other variables to register significant results include age (OR: 0.99, p < .05) and conservativism (OR: 1.39, p < .001). Contrary to the “cohort replacement hypothesis,” which presumes a liberalizing succession among younger generational cohorts, older respondents are more likely to withhold their punitive support. As expected, however, conservatives are more likely to support “tough on debt” approaches. Once stratification beliefs are included in Model 1b, the overall Black-white effect is no longer statistically significant. Attributions of unpaid legal debt, particularly individualism, explain it away. Compared with structuralists, individualists are 161 percent (OR: 2.61, p < .001) more likely to support the restriction of housing assistance, all else equal. Age (OR: 1.39,p < .001) and conservativism (OR: 0.98, p < .01) remain salient with similar effect sizes. The AIC for Model 1b is 1,356, suggesting an improvement in fit relative to the base model (ΔAIC = 28).
Models 2a and 2b predict support for restricting access to government-provided nutritional supplements. Model 2a shows ethnoracial differences on two scores. Blacks are 57 percent (OR: 0.43, p < .001) less likely than whites to support this policy, and Latinxs are 35 percent (OR: 0.65, p < .05) less likely than whites to extend their support. This Latinx-white difference is exceptional. In contrast to Blacks who consistently depart from whites on most every outcome considered, Model 2a is the only model where a significant difference exists between Latinxs and whites, suggesting ideological proximity between these two groups that may reflect a shifting colorline (Yancey 2003). The mediators in Model 2b reduce the Black-white OR by 13.95 percent (ΔOR = 0.06), but the effect remains significant (p < .001). Adding attributions of unpaid legal debt in Model 2b renders the Latinx-white OR statistically insignificant. Those who identify as individualists are 142 percent (OR: 2.42, p < .001) more likely than structuralists to support restricting access to food stamps. As with Models 1a and 1b, Models 2a and 2b show that age (Model 2a OR: 0.98, p < .001; Model 2b OR: 0.98, p < .001) and conservatism (Model 2a OR: 1.48, p < .001; Model 2b OR: 1.41, p < .001) are salient with effects in the same direction and of similar size. Unlike the previous models, however, education is also a salient predictor. Higher levels of education predict more liberal views on restricting access to food stamps (Model 2a OR: 0.86, p < .01; Model 2b OR: 0.87, p < .05). Change in the AIC values (ΔAIC = 21) shows improvement between Models 2a (AIC = 1,323) and 2b (AIC = 1,302).
Models 3a and 3b predict support for restricting access to government-assisted medical care. Model 3a shows that Blacks are 56 percent (OR: 0.44, p < .001) less likely than whites to support this policy, and this effect size reduces by 9.09 percent (ΔOR = 0.04) once we account for stratification beliefs in Model 3b. All else held constant, Blacks become 52 percent (OR: 0.48, p < .001) less likely than whites to support limiting Medicaid. Like our previous models, age (Model 3a OR: 0.98, p < .001; Model 3b OR: 0.97, p < .001) and conservatism (Model 3a OR: 1.44, p < .001; Model 3b OR: 1.40, p < .001) deliver significant results. The AIC value of 1,276 for Model 3b indicates that the inclusion of our mediators is an improvement over the fit of Model 3a, which has an AIC value of 1,290.
Models 4a and 4b predict support for restricting access to financial aid for education. Like the base models that precede it, Model 4a shows Blacks are less likely than whites to endorse the punitive response (OR: 0.48, p < .001). This effect size reduces by 12.50 percent (ΔOR = 0.06), but remains statistically significant (p < .01), when stratification beliefs are considered. Model 4b predicts that individualists are 128 percent (OR: 2.28, p < .001) more likely than structuralists to support the idea of restricting financial aid for those with unpaid legal debt. Unlike with our three other dependent variables (Models 1a–3b), age is not a salient predictor in Models 4a and 4b. Conservatism, though, is significant in both models (Model 4a OR: 1.33, p < .001; Model 4b OR: 1.28, p < .001). Conservatives are more likely to endorse punitive recourse. When we consider change in AIC values (ΔAIC = 22), it is safe to conclude that Model 4b (AIC = 1,425) is an improvement over Model 4a (AIC = 1,403).
Government Confiscation of Resources
The second set of questions ask respondents if they would support state confiscation of resources from those with unpaid legal debt under certain circumstances. We use binary logit models to predict support (1 = yes, 0 = no) for expropriation under circumstances that include when a debtor misses debt payments, is a recipient of government assistance, spends their money elsewhere, or becomes involved in illegal activity. These questions dovetail with a political moment where local governments are increasingly adopting consumerist policies that attempt to recoup the expenses of punishment, where “captive consumers” are billed at many pay-to-stay cost points (e.g., phone calls, video visitation, care packages, food and commissary) during incarceration (Harris, Smith, and Obara 2019). As with the previous set of questions, each dependent variable includes two models. All else equal, the base model determines if ethnoracial groups have different odds of supporting government confiscation under varying scenarios. The second model includes intervening measures of stratification beliefs in addition to the original variables featured in the base model. We include AIC values to show there is substantial improvement in efficiency between all our model pairs.
Results presented in Table 7 generally reflect the patterns we observe in our previous models, with three notable differences. To avoid redundancy, the analysis to follow emphasizes these points of departure. The first departure regards Black-white differences. These are more consistent predictors of support for restricting the social safety net than of predicting support for expropriation. Whereas seven of the eight models in Table 6 offer statistically significant results, only four of the eight models in Table 7 return significant ORs. There are no statistical differences between Blacks and whites on support for government confiscation when debtors either have spent their money elsewhere or were involved in illegal activity. When debtors miss debt payments, however, Blacks are 39 percent (OR: 0.61, p < .05) less likely than whites to support dispossession. Blacks are also 54 percent (OR: 0.46, p < .001) less likely than whites to support confiscation when debtors receive government assistance. These inconsistent findings between Models 5a to 6b, on one hand, and Models 7a to 8b, on the other, indicate mixed support for the expectations outlined in our first hypothesis.
Do Modes of Explanation Mediate Attitudinal Differences toward Government Confiscation of Resources for Outstanding Legal Debt under Certain Circumstances?
Note. The main entries are odds ratios. AIC = Akaike information criterion.
The reference category includes whites.
The reference category includes those who identify with the Democratic Party, Democratic Socialists, and the Green Party.
The reference category includes those classified as structuralists.
*p < .05. **p < .01. ***p < .001 (two-tailed tests).
The second point of departure involves crime salience factors. Whereas crime salience is not significant in our first set of models that predict limiting the social safety net, these variables become significant in two of our eight models predicting support for government dispossession. Our initial expectation (H3) led us to expect crime salience factors would correspond with support for punitive recourse, especially given how crime exposure can be so personally gripping. Instead, we find little evidence that these measures have predictive value. When they do render significant results in Table 7, a distinction appears to be between the types of questions asked. Crime salience becomes important under circumstances often refracted through a lens of racist stereotypes (e.g., welfare recipients, criminals). Model 6b predicts that respondents who sense crime is high in their neighborhood are more likely to support expropriation when these debtors are recipients of government assistance (OR: 1.23, p < .05). Model 8a predicts that respondents, or their family members, who have been crime victims are more likely to support dispossession when debtors are involved in illegal activity (OR: 1.41,p < .05). However, this effect is explained away once our mediators are considered (M8b).
Third, departing from all the other models is the effect size of individualism in Model 6b: “recipient of government assistance.” Whereas individualists are 215 percent (OR: 3.15, p < .001) more likely than structuralists to support expropriation if debtors receive government assistance, the other models report ORs that range from 1.66 (p < .05) in our “money spent elsewhere” model (M7b) to 1.83(p < .01) in our “involved in illegal activity” model (M8b). The difference in effect size is nearly twofold, perhaps because questions about welfare policy are often color-coded in their meaning. Other studies have shown that American opposition to welfare is largely driven by beliefs that equate public aid with specific minority groups, assume these groups abuse this support, and judge those on welfare as the undeserving poor (Gilens 1999). When our respondents were asked if they support expropriating the resources from those who owe legal debt and receive welfare benefits, it is plausible that this combination of questions conjures welfare resentment among respondents in ways that amplify the link between individualism and punitiveness.
Why it Matters: Understanding the Black-White Punishment Gap
In a sociohistorical moment where government operations like carceral punishment are increasingly defined by consumer-purchasing logics (Friedman 2021), our study shows that a complex relationship exists between ethnoracial position, stratification beliefs, and viewing unpaid legal debt as a breach in the social contract. Debt becomes a “negative credential” (Pager 2003) for some respondents, justifying punitive policy responses that exclude people from public assistance and forcibly take their possessions under certain circumstances. To summarize our results, we find differences in how people explain the origins of unpaid legal debt along key measures of socio-demographics and political values. We also identify differences, especially among Blacks and whites, in these attributions with other measures held constant. Ranging from individualist to systemic accounts, how people understand why legal debt goes unpaid shapes their policy response preferences, and they explain a small but significant fraction of ethnoracial differences in desires for more punitive recourse. These findings are consistent with much of the literature on stratification beliefs, but our study is unique in that it extends these insights to attitudes toward the escalation of monetary punishment.
The punishment gap on monetary sanctions between Blacks and whites, with Latinxs falling in-between, calls for more attention to be placed on the innerworkings between positionality and everyday experiences to unpack the circumstances under which stratification beliefs are applied as well as the meanings people assign to understanding why legal debt goes unpaid. We take the marginal intervening effect to mean that stratification beliefs are an important part of the broader narrative behind policy differences among Blacks and whites, but that these beliefs do not capture the entire story as the ethnoracial gap persists. It is our suspicion that patterns beyond the attributional process account for these Black-white differences in punishment preferences: namely, the differential exposure each group has with the justice system and the state generally.
For many people living in Chicago’s communities of color, the face of the justice system is among the most visible manifestations of state power. It may be the only face of government that some are able to recognize (Soss and Weaver 2017). Rather than participate in elections, demonstrate in Daley Plaza, or watch political debates on television, their most direct experiences with the state often come through interactions with police officers, probation and parole officials, county clerk administrators, and city payment portals, all of which labor in unique ways under the punishment machinery of monetary sanctions to collect government revenues. The substance of these relationships does not center on democratic values. For those with unpaid legal debt, even if they pose no serious threat to themselves or others, their relationship to the state centers on monied transactions that extract resources and limit freedoms (Harris 2016). Unpaid legal debt erodes the traditional terms of citizenship, making rights, protections, and affordances conditional for those who cannot or do not pay it off. What our findings reveal is that there are significant segments of Chicagoans who would prefer that these punitive relations intensified, and by extension, that the authorities of penal and welfare state become more intertwined. These segments support recourse that uses unpaid legal debt as grounds of civic failure to dislodge people from a social safety net and dispossess whatever resources these debtors are unlikely to possess, seemingly penalizing the precarious for their own precarity.
Footnotes
Appendix
Acknowledgements
Let us extend our thanks to the reviewers and Editor David Brunsma for their constructive comments on earlier iterations of the paper. We are grateful for all the help along the way. Any remaining shortcomings are our own.