Abstract
Although men’s incarceration rates have declined in the United States, women’s have stayed steady and even risen in some areas. At the same time, courts have increased their use of monetary sanctions, especially for low-level offenses. We propose that women’s incarceration trends can be partially explained by county dependence on monetary sanctions as a source of revenue. We suggest that monetary sanctions expose female defendants to processes that increase their likelihood of incarceration, especially in counties more reliant on monetary sanctions as a source of revenue, and where women’s poverty rates are high. Using data from Washington State, we find county dependence on monetary sanctions is positively associated with rates of women sentenced to incarceration. Although rural counties’ rates are higher, they depend on monetary sanctions no more than nonrural counties do.
Following several decades of persistent and exceptional growth, incarceration rates in the United States began trending downward around 2007 (Leigh 2020; Zeng 2017). These declines are certainly cause for optimism among legal scholars and criminologists, who have long noted not only that is such growth unsustainable (National Research Council 2014), but also that the consequences of mass incarceration are far reaching, negative, and predominantly borne by poor, racialized communities and individuals (Alexander 2010; Beckett 2018; Clear 2009).
However, reductions in state prison populations have occurred almost entirely among men (Sawyer 2018). White men’s incarceration rates declined nationally by 11 percent from 2000 to 2016, White women’s increased by 44 percent, and Black women remain twice as likely as White women to be incarcerated (Sentencing Project 2018). In some states, including Washington State—the focus of this study—the women’s prison population has continued to grow even as men’s prison populations have decreased, and these sex differences are even more pronounced when it comes to jail populations (Kang-Brown and Subramanian 2017; Sawyer 2018; Swavola, Riley, and Subramanian 2016).1 Although women’s incarceration rates remain low relative to men’s, these diverging patterns stress the importance of understanding the specific factors that explain why women are becoming disproportionately affected by criminal justice policies.
Largely missing from the research on women’s incarceration is the increasing use of monetary sanctions, also known as legal financial obligations (LFOs), as a form of punishment in the United States and its growing role in supplementing county budgets (Edwards 2020; Martin 2018).2 Although LFOs are generally touted as a kinder alternative to incarceration, research indicates that people sentenced to LFO debt suffer greatly for it (Harris 2016; Martin et al. 2018). The use of monetary sanctions to supplement county budgets and pay for local criminal legal systems shifts the literal costs of doing justice from courts to defendants (Martin et al. 2018) and incentivizes increased policing of minor, misdemeanor offenses (Kohler-Hausmann 2013; Martin 2018). As local jurisdictions become increasingly dependent on LFO revenue, it becomes even more important to understand the potential impact on criminal justice outcomes.
There is reason to believe women have been especially hard hit by this shift. Although overall poverty rates have decreased substantially since the 1970s, women remain disproportionately represented among America’s poor (McLanahan and Kelly 2006; Wagner 2019). People who are poor are disproportionately represented in the criminal legal system in general and among those carrying LFOs debt in particular (Bing, Pettit, and Slavinski 2022, this volume; Harris, Evans, and Beckett 2011; Hunt and Nichol 2017; Peterson, Krivo, and Harris 2000). Given women’s economically marginalized position, they may be more likely to request incarceration over LFOs during plea agreements or sentencing arrangements, more likely to default on payments, and more likely to remain under the supervision of the criminal legal system for extended periods of time. We may therefore expect counties’ dependence on monetary sanctions to be associated with an increase in women sentenced to incarceration because women may be more vulnerable to carceral sentencing as a result of their real or perceived inability to pay monetary sanctions.
Finally, there is reason to believe that the relationship between LFO revenue and women’s incarceration will be stronger in rural counties. Across the United States, rates of women’s jail incarceration in small counties have nearly doubled since 1970, but have remained relatively stable in mid-size and large counties (McCoy and Russo 2018). Reports produced by criminal justice nonprofits and research institutions have argued that upticks in arrests for misdemeanor offenses and economic marginalization are key drivers of women’s incarceration in these spaces (Kajstura 2017; Kang-Brown and Subramanian 2017; McCoy and Russo 2018). One potential reason is that small and rural jurisdictions sentence defendants to monetary sanctions at higher rates and with larger amounts than nonrural jurisdictions (Olson and Ramker 2001; Ruback and Clark 2011). Given high poverty rates and growing jail populations, these increases are likely associated with rural jurisdictions’ need to pay for growing carceral and supervisory systems. Thus the system of monetary sanctions places rural counties in the precarious position of pursuing justice and revenue at the same time, potentially undermining the stated goals of the criminal legal system (Martin 2018).
Considering these factors, we explore whether women across the thirty-nine counties in Washington State are sentenced to incarceration at higher rates in counties that depend more on monetary sanctions as a source of revenue. We also test whether these associations are especially strong in rural counties. We combine data from 2007 to 2012 from the Washington State Administrative Office of the Courts (AOC), the Washington State Auditor’s Office, the Washington State Governor’s Office of Indian Affairs, and the U.S. Census Bureau to create a unique dataset used to trace associations among county dependence on monetary sanctions, women’s sentencing rates, and county characteristics. We then explore the effect of rurality on monetary sanction dependence and rates of women sentenced to incarceration, and test whether women’s incarceration rates may be explained by a county’s reliance on LFOs. We find that as the percentage of county revenue from fines and fees increases, so do rates at which counties sentence women to incarceration. We find rural counties are no more dependent on monetary sanctions than are nonrural counties, though they do sentence women to incarceration at higher rates than nonrural counties. Overall, our study indicates that rurality and county dependence operate as independent determinants of women’s incarceration sentencing rates with potentially compounding effects, but we do not find an interactive effect.
MONETARY SANCTIONS AND WOMEN’S INCARCERATION
Monetary sanctions are an increasingly common and widespread form of punishment at all levels of government and are imposed for all manner of cases, including citations, traffic violations, misdemeanors, and felony charges (Harris, Pattillo, and Sykes 2022, this volume). Although it is difficult to estimate the number of individuals who have outstanding legal debt, a report by the National Center for Victims of Crime (2011) estimates approximately ten million people across the country owe more than $50 billion in restitution alone, and these already striking figures do not capture debtors or debt attributable to fines, fees, or surcharges associated with criminal legal contact. The rising use of monetary sanctions has been linked to an increased dependence on fines and fees for court revenue (Fernandes et al. 2019; Martin et al. 2018). Karin Martin (2020) links this change to the rising cost of criminal justice intervention and the fiscal pressures put on local governments by the reduction of state budgets. As jurisdictions across the country scrambled to meet the financial burden of ballooning criminal legal systems, the turn toward the collection of LFOs seemed like a straightforward means of financial solvency. While recent research has demonstrated the inefficiency of LFO collection systems and the negative consequences that these practices can have on indigent individuals, a majority of jurisdictions use monetary sanction revenue as a portion of their general funds (Fernandes et al. 2019; Menendez et al. 2019).
Research on the system of monetary sanctions consistently finds it disproportionately impacts economically marginalized people. In their foundational work on monetary sanctions, Alexes Harris, Heather Evans, and Katherine Beckett (2010) demonstrate how LFOs increase social inequality by creating long-term debt for individuals who are unable to pay. Subsequent research has continued to outline the various collateral consequences on people who are poor, including housing instability, the loss of driving and voting rights, continued entanglement in the criminal legal system, and damage to one’s credit (Cadigan and Kirk 2020; Colgan 2019; Link, Hyatt, and Ruhland 2020; Pattillo et al. 2022, this volume). Thus the detrimental effects of LFOs on economically marginalized individuals and communities has become increasingly clear, an understanding reiterated by a number of articles in this volume (Bing, Pettit, and Slavinski 2022; Boches et al. 2022; Harris and Smith 2022; Pattillo et al. 2022; and Sykes et al. 2022).
Unfortunately, few studies have examined whether monetary sanctions have differential impacts on women, though this may be the case.3 First, women are more likely to be represented in cases involving lower-level offenses and it is these types of offenses for which LFOs are the predominant form of punishment. Lindsay Bing, Becky Pettit, and Ilya Slavinski (2022, this volume) demonstrate this directly, showing that women made up between 10 and 14 percent more of the fine-only misdemeanor cases in Texas than cases involving more serious offenses. This means policy changes that increase policing of misdemeanor offenses are especially impactful on women (Chesney-Lind and Pasko 2013; Schwartz 2013), and attempts to shift sentencing toward noncarceral punishments likely increase the number of women carrying legal debt. Second, the financial hardships created and exacerbated by legal debt may be especially pronounced among women because women are more likely to experience economic disadvantage than men (McLanahan and Kelly 2006; Wagner 2019). If counties are increasingly relying on monetary sanctions to fund their legal systems, or are more likely to impose monetary sanctions as a form of punishment, then the economic circumstances of women become increasingly important for determining sentencing outcomes.
The fact of women’s economic precarity can help us theorize why LFO sentencing rates would be related to incarceration sentencing rates. First, women may be more likely to opt for short stints of incarceration than for legal debt. If they are unable to afford the LFOs they would accrue, then they may seek to serve their punishment without the associated financial burden. Second, women may be more likely to default on legal debt and be incarcerated for nonpayment. Finally, the imposition of LFOs would increase their interaction with the courts and leave them “tethered” to the criminal justice system (Harris 2016). This tethering may make women debtors more visible to law enforcement, increasing their risk of arrest, even as their newfound status as “repeat offenders” places them at higher risk of receiving a carceral sentence. Thus inability to pay legal debt is likely one of many mechanisms through which low-income women find themselves disproportionately represented among those involved with the criminal justice system. We cannot observe these mechanisms directly with our data, but the plausibility of these situations motivated us to examine the linkage between the use of LFOs as a form of punishment and the number of women sentenced to incarceration.
SPATIAL TRENDS IN MONETARY SANCTIONING
Recent research on monetary sanctioning has stressed the importance of analyzing the consequences of monetary sanction dependence as a spatially contingent phenomenon (DOJ 2015; Edwards 2020). Studies on the spatial determinants and correlates of monetary sanctions provide a roadmap to understanding the distribution of legal debt across the United States. Research by Alexes Harris and her colleagues (Harris 2016; Harris, Evans, and Beckett 2011) suggests that particular counties and ethnic groups are disproportionately saddled with LFO debt. Additional reporting from Governing magazine finds that small towns and rural areas across the United States especially depend on fines and fees as a source of county revenue, some jurisdictions reporting upward of 90 percent of their general revenues in 2017 and 2018 as having been generated by monetary sanctions (Maciag 2019). Finally, Gabriela Kirk and her colleagues (2022, this volume), show how relationships between defendants and court actors in small jurisdictions can impact court decision-making. They found court actors in rural settings often knew defendants more personally and were keenly aware of their financial precarity, but that the same court actors felt restricted in their ability to apply discretion in sentencing because of LFO laws and policies.
These studies, as well as broader examinations of LFO statutes, demonstrate the “decentralized, poorly coordinated, and inconsistent” nature of LFO practices across jurisdictions and the ways in which they “concentrate negative impacts among people with low incomes” (Friedman et al. 2022, this volume). Given that court systems are organized by geographic area, dependence on monetary sanctions and what it means for both individual and county-level criminal justice outcomes cannot be wholly disentangled from the structural characteristics of the spaces in which the system of monetary sanctions operates.
These localized contingencies are especially important when we consider their probable impact on women’s incarceration rates. Although literature on the spatial determinants and correlates of monetary sanctions is limited, sociological and criminological literature on the spatial distribution of incarceration and criminal justice contact consistently find location to be an important predictor of carceral trends (Clear 2009; Sampson 2012). Criminologists who study trends in women’s justice involvement in particular have long argued that structural explanations of women’s offending, arrest, and incarceration explain how these phenomena are patterned across time and space more effectively than do behavioral explanations (Chesney-Lind and Pasko 2013; Schwartz 2013; Steffensmeier and Allan 1996). In addition to being leaders in LFO revenue generation, rural counties are also implicated in maintaining and even increasing women’s incarceration rates in the face of men’s decelerating rates (McCoy and Russo 2018; Sentencing Project 2019). Feminist scholars have focused broadly on how and why tough-on-crime policies have increased women’s visibility and representation in the criminal legal system, but few have explored how jurisdictional attempts to pay for the rapid expansion of these systems may themselves be tied to increases in women’s criminal justice contact. Given the demonstrated economic precarity of women in rural areas (Snyder and McLaughlin 2004), one might expect these factors combine to make it more difficult for women in rural areas to pay off monetary sanctions than women in nonrural areas, and more vulnerable to incarceration.
CURRENT STUDY
Overall, we suggest that an association between recent trends in women’s incarceration rates (National Research Council 2014), the emergence of monetary sanctions as a sentencing option (Harris, Evans, and Beckett 2010), and rural areas’ alleged higher dependence on monetary sanctions as a source of revenue (Olson and Ramker 2001; Ruback and Clark 2011). In counties where the system of monetary sanctions is a significant revenue generator, LFOs may take on a primary role in punishment. This increased use of legal debt may leave women particularly vulnerable to circumstances that increase their likelihood of arrest and incarceration. These factors coalesce to increase rates of women arrested, tried, and sentenced to incarceration while women’s economically marginalized position makes it harder for them to pay off their legal debts and remove themselves from criminal legal supervision. We expect that as county dependence on monetary sanctions increases, so will rates of women sentenced to incarceration. Further, because these factors are likely stronger in smaller, poorer jurisdictions, we expect this association to be stronger in rural counties.
DATA
We calculate county-level women’s incarceration sentencing rates between 2007 and 2012 using data from the Washington Administrative Office of the Courts.4 The AOC is responsible for the supervision and information management for every court jurisdiction within the state. The dataset includes case and defendant information for thirty-nine superior courts, sixty-one district courts, and ninety-seven municipal courts.5 Each case has a coded indicator of whether incarceration was imposed during sentencing. In superior court cases, a dichotomous identifier indicates whether an individual was sentenced to prison or jail. For district and municipal courts, codes indicate whether the defendant was sentenced to jail and the number of days that were sentenced. We used these measures to generate two dichotomous variables tracking whether defendants were sentenced to jail (365 or fewer days of incarceration) or prison (366 or more days). We then summed affirmative cases within counties, by year, and generated women’s annual jail and prison sentencing rates per every hundred thousand adult women.
Although most cases had gender information for defendants, this information was unavailable for a subset of cases.6 For these, we imputed defendant gender based on people’s first names, using the gender package in the R programming language. This package uses birth year and first name to find the proportion of people assigned male or female at birth with that name in that year. It then imputes a gender designation based on this information. The process was 98 percent accurate when this information was available, and results were substantively identical in models comparing imputed versus provided gender designation data.
Information on county budgets came from public records generated by the Office of the Washington State Auditor, which is responsible for overseeing the collection and usage of state and local government funds. Information collected by the state auditor’s office includes an accounting of all revenues and expenditures for each county government within a given fiscal year. This data indicated the total amount of county revenue generated through the collection of financial penalties, including court-ordered obligations from both criminal and traffic offenses. We calculated the annual percentages of county revenue derived from fines, fees, and penalties by dividing revenue generated from these penalties by the total county revenue and multiplying the resulting proportion by one hundred.7 This gave us a measure of how much a county depended on monetary sanctions for revenue generation each year, relative to other revenue sources. Given county dependence on monetary sanctions may exert both an immediate and lagged effect we control for both contemporaneous and prior (t-1) dependence on monetary sanctions in our final models.
Demographic and population information for each county comes from the Census Bureau. Annual population estimates are provided by the Census Bureau’s Population Estimates Program, which uses current data on births, deaths, and migration to calculate population change since the most recent decennial census for federal, state, county, and local entities.8 Demographic information on the gender, age, race, and poverty composition of each county are fixed estimates pulled from the 2010 Decennial Census. Given our focus on adult incarceration sentencing rates, we exclude county residents under the age of eighteen from our sentencing rate calculations. We estimate the number of adult female residents by multiplying the estimated total population of the county in a given year by the proportion of the population that is female, and then by the proportion of the population age eighteen or older. We also created a dichotomous variable to indicate whether the county is considered a rural county according to the standards set by the Washington State Office of Financial Management. In accordance with this classification, counties whose population densities were fewer than one hundred people per square mile, or that were smaller than 225 square miles, were marked as rural in our analyses. One county saw its status change from rural to nonrural halfway through the six-year data collection period given a gradual increase in its population density. All other counties were consistently rural or nonrural throughout collection.
The racial and ethnic composition of Washington State’s rural counties differ from national norms and suggest the inclusion of separate controls for race and ethnicity. Nationally, 89 percent of rural residents are White, non-Hispanic (Parker et al. 2018). Only 76 percent of rural Washingtonians in our analyses are White, however, in spite of the fact that Washington is on average slightly whiter than the rest of the country. An above-average portion of Washington’s non-White population—particularly non-White Hispanic persons—live in rural areas; of the ten counties in Washington with the largest Hispanic populations eight are rural, and two of those rural counties are majority Hispanic. Further, the demographics of Washingtonians of color differ substantially from national averages, particularly in regard to its below-average proportion of Black residents and above-average proportion of Asian residents. Given that the legal system in the United States produces unequal outcomes based on race and ethnicity, counties likely vary in terms of their dependence on monetary sanctions and rates of women sentenced to incarceration depending on the extent to which their populations are racialized. We therefore include both the percentage of people who are non-White in the census, regardless of ethnic identification, and the percentage of people who are Hispanic, regardless of racial identification, as two separate covariates in our models. In doing so, we account for both rural Washington’s unique demographic composition and the discriminatory conditions experienced by non-White and Hispanic residents.
Finally, we include two controls for varying jurisdictional arrangements within counties. First, Washington State contains the seventh largest Native American population in the United States and about half of all Washington State counties include federally recognized Native American reservations, officially referred to as Indian Country by state.9 Because Native Americans are overrepresented among those who experience criminal justice contact (Males 2014; Nielsen and Silverman 1996), and sovereign reservations often maintain their own judicial systems, we considered it important to capture the dynamics between Native American populations, Tribal Court sovereignty, and sentencing rates. We generated a fixed, dichotomous variable measuring whether each county contains Indian Country based on information from the Washington State Governor’s Office of Indian Affairs. This measure includes twenty-nine federally recognized tribes and their land.10
As a second jurisdictional control, we include a measure of the number of municipal courts (sometimes called city courts) within a county. Research has recognized the tendency of municipal courts in particular to focus on generating revenues through legal fines and fees imposed for low-level or misdemeanor crimes (Fernandes et al. 2019). Although we include municipal-level cases in our sentencing rate, we recognize that these courts may have a heavier influence on the relationship between revenue generation and sentencing rates than other court levels. We therefore include a fixed count of the number of municipal courts operating within each county as indicated by the Washington State Court Directory.11 Notably, the population size of a municipality and number of municipal courts are inextricably linked: larger cities and towns have more municipal courts. As a result, our municipal court count is strongly associated with our measure of county population (α = 0.82), and moderately associated with our measure of rurality (α = –0.46): These associations are likely because as county population increases so does the number of municipal courts, and rural counties’ populations are generally smaller than those of nonrural counties.
METHODS
We posed three hypotheses in our analysis to explore the associations between rurality, female sentencing and incarceration, and monetary sanctions.
H1: Rurality is positively associated with county dependence on monetary sanctions.
We test the first hypothesis using a random-effects model. The random-effects model is more flexible than the fixed-effects in that it allows us to explore the potential effects of time-invariant variables such as percentage of county residents in poverty. We use county-specific random intercepts to account for the observation interdependence within counties over time. These intercepts include a county-specific error term that is assumed to be uncorrelated to the other covariates in the model. Effects were estimated using generalized least squares.
H2: Rurality is positively associated with female incarceration sentencing rates.
H3: County dependence on monetary sanctions is positively associated with female sentencing rates.
We test our second and third hypotheses using random-effects negative binomial models. As with the first hypothesis, we use random effects to control for observation interdependence. In addition, most counties’ annual female sentencing rates are low, and we find that our measure of counties’ annual female sentencing rates overdispersed (the variance exceeds the mean), indicating OLS and Poisson models will produce biased coefficient estimates. Therefore, we use negative binomial models, which include a parameter to account for this overdispersion and are designed to model count outcomes, such as a count of how many women per hundred thousand in the population are sentenced to incarceration each year. Given strong associations between many of our control variables with county population size, we exclude county population from our final negative binomial models.
RESULTS
Results generally support our hypothesis regarding the relationship between women’s carceral sentencing and county dependence on monetary sanctions. We also find, however, that rurality has little effect on monetary sanction dependence, despite its association with women’s incarceration.
RELATIONSHIPS AMONG RURALITY, COUNTY DEPENDENCE, AND SENTENCING
Our first hypothesis explores the relationships between rurality, sentencing, and county dependence on monetary sanctions.
As table 2 shows, our models show no significant difference between dependence on monetary sanctions in rural and nonrural counties. Notably, although rurality and the percentage of residents in poverty is fairly strongly correlated (α = 0.60 in 2012), the effect of rurality is insignificant regardless of the inclusion of the percent in poverty variable.12 A 1 percent increase in residents in poverty, however, is associated with a 0.10 percent increase in county dependence on monetary sanctions. These findings indicate dependence on monetary sanctions may be more closely related to poverty than to rurality, suggesting economic motivations for counties’ turn to fines and fees as a source of county revenue.
As table 3 shows, models support our second hypothesis, that rurality is positively associated with female incarceration. We find that county rurality is associated with a 23 to 27 percent increase in rates of women sentenced to incarceration in models 1 and 2.13
In addition to supporting the second hypothesis, findings align with previous research and arguments on geographic determinants of incarceration rates across the United States (McCoy and Russo 2018; Sentencing Project 2018). In fact, all county characteristics included in models 1 and 2 are associated, to varying degrees, with rates of women sentenced to incarceration.
The third hypothesis, that county dependence on monetary sanctions is positively associated with female incarceration sentencing rates, is supported for both contemporaneous county dependence on monetary sanctions and the prior year’s county dependence on monetary sanctions.14 Model 1 includes a contemporaneous measure of county dependence on monetary sanctions; model 2 includes both contemporaneous and the prior year’s dependence on monetary sanctions. In model 1, a 1 percent increase in the percentage of county revenue derived from monetary sanctions is associated with a 23 percent increase in rates of women sentenced to incarceration. In model 2, we look at both current and prior dependence on monetary sanctions, and find that a 1 percent increase in current county dependence is associated with an 18 percent increase in rates of women sentenced to incarceration. The prior year’s dependence on monetary sanctions is also associated with increases in rates of women sentenced to incarceration: a 1 percent change in the previous year’s county dependence on monetary sanctions is associated with a 12 percent increase in sentencing rates. These results support the third hypothesis in that county dependence on monetary sanctions is positively associated with rates of women sentenced to incarceration.
Research on socioeconomic disparities in incarceration leads us to expect that counties with higher poverty should sentence women to incarceration at higher rates, all else equal. However, our regression results showed a negative coefficient for the relationship between poverty and women’s incarceration sentencing. This finding was initially puzzling, but we believe that it reflects the complex relationship between county dependence on monetary sanctions and poverty. We operationalize county dependence as a measure of the percentage of LFOs collected by counties within a given fiscal year relative to overall revenue. Unlike the sanctioning of LFOs, the collection of LFOs is directly related to the economic circumstances of individuals who experience such sanctions. Considering this, we would expect counties with higher poverty rates to collect a smaller proportion of sentenced LFO payments than more affluent counties given residents’ inability to pay. Because our model shows a strong positive effect between reliance on monetary sanctions and rates of women sentenced to incarceration, it may be that the effect of poverty is captured by our county dependence variable. However, it is always possible that the unique demographic or legal characteristics of Washington could explain these unexpected findings. Teasing out such associations is beyond the scope of this article.15 That said, we leave poverty in our model as a control in order to capture any potential residual effects and encourage future researchers to more formally test the hypotheses implied by our results.
A second result standing in opposition to findings in earlier studies is that the percentage of non-White residents per county is associated with decreases in rates of women sentenced to incarceration. However, a direct interpretation of our race and ethnicity coefficients is complicated by their strong relationships with one another. Our results imply that, all else equal, women’s incarceration rates are slightly lower in counties with larger percentages of people who are racialized as Black, indigenous, people of color (BIPOC). However, changes in the percentage of BIPOC residents and percent of Hispanic residents both imply changes in percentages of people racialized as BIPOC. Still, although the effect of the percentage of BIPOC residents is small in magnitude, it is statistically significant and certainly surprising.
INDIAN COUNTRY, COUNTY DEPENDENCE ON MONETARY SANCTIONS, AND SENTENCING
In the course of validating our hypotheses, a clear pattern regarding Indian Country, monetary sanctions, and sentencing emerged. Counties that contain Indian Country have significantly higher sentencing rates than those without Indian Country. The magnitude of these effects varies slightly across models, but on average Indian Country is associated with rates of women’s incarceration being 40 percent higher than other areas of the state. Further, during our study period counties containing Indian Country generated less revenue from monetary sanctions than other counties did. Given the presence of sovereign criminal legal systems and courts on many Native American reservations across the United States, this result makes sense. Counties containing Indian Country have portions of their offending populations ineligible for processing in the courts captured in these analyses and thus are unable to collect monetary sanctions from defendants processed in tribal courts. However, the finding that rates of women sentenced to incarceration are actually greater in counties containing Indian Country cannot be explained away by referencing the mechanics of tribal versus nontribal courts. We had anticipated that counties containing Indian Country would have lower sentencing rates because portions of their populations are ineligible for sentencing in the courts captured in these analyses. Instead, we find the opposite: counties containing Indian Country sentence women to incarceration at higher rates than do counties that do not.
DISCUSSION AND CONCLUSION
Our results highlight the importance of exploring the underlying mechanisms that drive trends in women’s incarceration and how they relate to the increasing use of monetary sanctions as a form of punishment. We find that the more reliant counties in Washington State are on monetary sanction revenue, the higher their rate of sentencing women to incarceration. This study thus takes the first step in establishing a relationship between monetary sanction revenue and the punishment of women. Still, the relationship between monetary sanctions, local criminal legal practices, and the economic marginalization of women is multifaceted, and more work is needed to disentangle these complex relationships. One possible explanation for the association between county dependence on monetary sanctions and women’s incarceration is the increased policing of low-level crimes. Low-level offenses have always made up a larger share of women’s offending than men’s (Chesney-Lind and Pasko 2013). Also, women may be especially likely to be swept into the system in counties where revenues generated off the policing of these low-level offenses are used to keep penal systems afloat. Our findings may also be explained by what Franklin Zimring and Gordon Hawkins (1991) call “the correctional free lunch.” From this perspective, it may be that counties more reliant on monetary sanctions are more likely to sentence women to incarceration as a way to shift the cost burden of punishment onto the state. Of course, future studies will need to examine these potential links more concretely. Specifically, qualitative research on jurisdictional policing and judicial practices could shed light on how and why the system of monetary sanctions influences such practices. Statistical work could also be done to examine the magnitude of low-level offenses in moderating this relationship.
Our findings in rural counties present a more nuanced explanation of women’s carceral trends than those suggested in much of the criminal justice reports and academic research. Although rurality is indeed associated with higher rates of women sentenced to incarceration, it is not associated with higher levels of dependence on monetary sanctions as a source of county revenue. It appears instead that dependence on monetary sanctions and rurality operate independently of one another to increase rates of women sentenced to incarceration across the state. This suggests that theories linking rural jurisdictions directly to revenue-generating policing and punishment practices are overlooking important explanatory factors. Again, the economic marginalization of women in rural spaces may explain why women in these areas are sentenced to incarceration at higher rates than their counterparts in nonrural areas (Barnett and Mencken 2002; McLanahan and Kelly 2006; Snyder and McLaughlin 2004). Although we did not find an association at the sentencing level, it may still be possible that LFOs function as a bridge over which the reduced economic means of rural women leads to incarceration. Research at the individual level is needed to validate this proposed relationship. We recommend that future researchers explore the relationship between individual women’s debt and likelihood of incarceration for infractions tied to its nonpayment and increased supervision.
These findings raise several directions of exploration in addition to those discussed. One pressing question is whether the relationship between rurality and dependence on monetary sanctions for revenue operates at the city or town level rather than county level. A recent analysis by Governing magazine of city and town budgets in 2017 and 2018 identified three localities in Washington State that credited more than 10 percent of their general revenues to monetary sanctions (Maciag 2019). In our analyses, no one county credited more than 5.25 percent of its annual revenues to monetary sanction collections, suggesting variations in measurement and jurisdictional focus may influence results and obfuscate the complicated relationship between monetary sanction dependence and carceral trends. Further research is needed to determine the true relationship between rurality and local dependence on monetary sanctions. We recommend that future researchers explore this relationship at the city and town level.
Our findings regarding sovereign Native American territories and women’s sentencing rates suggest that intersections of race and space shape the consequences of monetary sanction dependence on carceral trends. Although counties containing Indian Country rely less on LFOs for revenue, they tended to incarcerate women at a higher rate. This finding is particularly striking given that tribal courts likely remove many Native American women from consideration for sentencing in the data used in these analyses. Somewhat counterintuitively, we find higher rates of women being incarcerated overall in counties where a segment of women are not under the jurisdiction of state courts. These findings support the study by Robert Stewart and his colleagues (2022, this volume) who find that Native Americans in Minnesota were subject to larger amounts of LFOs than White individuals and that these effects were even greater when courts were near tribal lands. A more direct and in-depth analysis of LFOs as they relate to tribal court sovereignty and native populations is necessary.16
Finally, although our study does not directly implicate sentencing or policing to generate revenue, it clearly points to a link between monetary sanctions and sentencing outcomes for women. We agree with researchers that the use of monetary sanctions as a source of jurisdictional revenue creates tension with the stated purpose of meting out justice (Martin 2020). We therefore recommend that counties explore streams of funding unassociated with the penalization of its residents to maintain their criminal legal systems and that they take steps to reduce the costs of implementing justice. Courts need to understand and address the ramifications and downstream consequences of monetary sanctions on individuals—especially women—who have a limited ability to pay. Even sentencing policies that appear lenient can lead to incarceration whenever LFOs are involved, and a lack of data on defendants sentenced to incarceration for willful nonpayment prevents a full accounting of the scope of this social problem.17
Our findings complicate assertions that rural counties are both more punitive and more dependent on monetary sanctions as a source of revenue in that we find support only for the former. Nonetheless, our results suggest county dependence on monetary sanctions is one of several drivers of women’s sentencing outcomes across Washington State. The finding that monetary sanctions—a criminal legal punishment often touted as a kinder alternative to incarceration—may in fact be associated with increases in incarceration is alarming indeed. If the system of monetary sanctions is meant to act as a mechanism to recover the costs of the criminal legal system through collections and to provide an alternative to the seemingly more costly option of incarceration, then it has failed on both counts. This research contributes to a growing body of literature demonstrating that the primary goals of the system of monetary sanctions are being subverted by the system itself. The proliferate use of monetary sanctions does not reduce rates of women sentenced to incarceration, and counties committed to reducing their carceral populations can and should consider alternative revenue sources and sentencing options for women defendants.
FOOTNOTES
↵1. We use the term sex differences in our discussion of carceral trends because determinations regarding where to house people sentenced to incarceration are generally made using biological sex (genital anatomy) criteria, rather than gender identity criteria (National Center for Transgender Equality 2018).
↵2. As defined in the introduction to this volume, LFOs refer to “fines, fees, costs, restitution, surcharges, and other financial penalties that are imposed on people when they encounter the United States criminal legal system” (Harris, Pattillo, and Sykes 2022).
↵3. Daniel Boches and his colleagues (2022, this volume) provide an interesting examination of how LFOs affect family members indirectly. They point out that female partners and mothers of men who are imposed LFOs are often coerced into paying the legal debt themselves. These arrangements can cause serious tension in familial relationships.
↵4. For more details about the collection of administrative data within Washington, and across all of the sites in the larger study, see the introduction to this volume (Harris, Pattillo, and Sykes 2022). Our primary reason for selecting Washington as our case study was that it was one of two states that contained full sentencing data across a significant time period.
↵5. Unfortunately, the data from the Washington AOC does not include the Seattle Municipal court system, which maintains its own separate database. We recognize that the omission of such a large municipal district could bias our results. We therefore ran separate models that excluded the other courts in King County as a robustness check. Differences in coefficients between the models were nominal and the substantive impact on our variables of interest was unchanged. This suggests that mechanisms operating within King County are similar in effect to those operating in other counties, and that the omission of Seattle Municipal data does not substantially affect our findings.
↵6. We use the term gender here because our data may be more reflective of individual gender identities than biological sex assigned at birth (National Center for Transgender 2018). Washington State lists “sex” (male, female, or X) on its driver’s licenses, but allows residents to request a “change of gender designation” and display their gender identity on their license in the “sex” field. In addition, residents can change their names and sex designations with the Social Security Administration to correct their records and align them with their gender identity.
↵7. Total county revenue was the sum of all revenue from fines and penalties, taxes, license and permits, intergovernmental grants, usage fees, rents and leases, investment interest, and other miscellaneous sources.
↵8. Census Bureau, “Population and Housing Unit Estimates,” https://www.census.gov/programs-surveys/popest.html (accessed August 19, 2021).
↵9. World Atlas, “States with the Largest Native American Populations,” https://www.worldatlas.com/articles/us-states-with-the-largest-native-american-populations.html (accessed August 19, 2021).
↵10. Five of the thirty-four tribes in Washington State are not eligible for inclusion in these analyses because they are not federally recognized (see University of Washington 2020).
↵11. Washington Courts, “Washington State Court Directories,” https://www.courts.wa.gov/court_dir (accessed August 19, 2021).
↵12. Notably, rural × percent in poverty interaction terms were statistically insignificant in supplementary analyses.
↵13. For interpretation, we convert negative binomial coefficients into percent changes in-text, using the following formula: (expβ-1)×100=percent change in female sentencing rates
↵14. In addition to main effects in both models, we tested interactions between our controls and our variables of interest. None of the tested interactions were significant and the inclusion of interactions did not affect the magnitude of our main effects. We also tested a measure of income inequality as an addition control but found that its inclusion did not improve model fit. The interaction between rurality and income inequality was significant during these additional tests. This may indicate that although generally places with more income equality tend to have fewer women incarcerated, that association is essentially not present in rural areas. However, because it did not contribute to explaining the variance in our model, we chose not to include it.
↵15. Previous research has explored the association between macro-economic factors and monetary sanctions. Frank Edwards (2020) finds differences in the collection of LFOs by courts before and after the Great Recession of 2008. However, he did not find a connection between local finances and low-level criminal justice debt. Instead, differences were explained by increased caseloads and that “policing, traffic enforcement, and prosecutorial decision-making are likely the drivers of these shifts.” This study demonstrates how criminal justice outcomes vary not only by the individual characteristics of defendants, but also by physical location, timing, and jurisdictional characteristics.
↵16. Such studies should also build on the growing literature related to settler colonialism, as Stewart and his colleagues discuss: “Historically, settler governments extracted resources from tribal nations through coercion, displacement, and assimilation; today, these same entities draw financial resources disproportionately from American Indians through the criminal legal system. Put another way, settler colonial domination has transitioned from collective to individualized extraction from Native subjects, and thus the structure of settler colonialism persists” (153).
↵17. For a thorough qualitative exploration of the notion of willful nonpayment, see Fernandes, Friedman, and Kirk (2022, this volume).
- © 2022 Russell Sage Foundation. O’Neill, Kate K., Tyler Smith, and Ian Kennedy. 2022. “County Dependence on Monetary Sanctions: Implications for Women’s Incarceration.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8(2): 157–72. DOI: 10.7758/RSF.2022.8.2.08. We would like to thank Arnold Ventures for providing funding for this project (PI: Alexes Harris) and the Multi-State Monetary Sanctions Study team and Research Assistants who contributed to the data collection processes. Partial support for this research came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant, P2C HD042828, to the Center for Studies in Demography & Ecology at the University of Washington. Direct correspondence to: Kate K. O’Neill, at oneillkk{at}uw.edu, 211 Savery Hall, Box 353340, Seattle, WA 98195-340, United States.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.