Who Pays? Measuring Differences in the Process of Repayment of Legal Financial Obligations

Abstract

This study identifies the correlates of legal financial obligation (LFO) debt repayment among persons sentenced to probation and transferred to a specialized collections unit. Using bivariate tests and logistic regression, results indicate that starting balance amounts, monthly payment amounts, and enforcement actions (capias warrant) are the strongest influences on the likelihood of full debt repayment. These results indicate that some persons will struggle to repay their LFO balances if amounts assessed are in excess of their means, even in an institutional context adopting an individualized, flexible, and non-punitive approach to collections. Policy implications suggest a need for reform at the point of LFO assessment to avoid imposing obligations that are unreasonable to individuals’ ability to repay.

1. Introduction

Involvement with the criminal justice system can come at steep social and economic costs through ensuing collateral consequences (Kirk and Wakefield 2018) and, increasingly, financial debt (Martin et al. 2018). The latter results from the assessment of legal financial obligations (LFOs, hereafter), a phrase used here to refer to the numerous types of financial penalties that can be incurred throughout formal involvement with the justice system. LFOs, which can include fines, fees, costs, restitution, and surcharges, have been used increasingly in recent years (Menendez et al. 2019). Justice-involved persons can amass sizable LFO debt burdens throughout the entirety of criminal justice processing and punishment (Harris 2016; Harris et al. 2010; Link 2019; Martin et al. 2018; Pleggenkuhle 2018).

There has, however, been an imbalance of empirical attention: while many studies document the existence of LFO debt, fewer examine the extent to which assessed LFOs are repaid and the policies and infrastructures involved in the collections process. This gap in understanding exists amid acknowledgement on two key fronts. First, at the individual level, many lack the financial means to repay their LFO debts because there is an overrepresentation of people from socioeconomically disadvantaged backgrounds in the justice system (Harris 2016). Second, at the institutional level, LFO revenue is increasingly used as an essential source of funding for contemporary criminal justice systems (Fernandes et al. 2019). Accordingly, those facing challenges in paying assessed LFOs may be at risk of legal and/or social consequences for non-payment, especially in jurisdictions with a high reliance on this revenue (Link et al. 2020). Placing the responsibility of assessment and collection at the same level of government with dependence on this revenue for financing their operations can create perverse incentives for assessment and collections (Atkinson 2016). Sobol (2016) argued that LFO assessment and repayment creates a “two-tiered” system in which disadvantaged persons who are unable to pay LFO debt become entrapped in the system and cycles of punitive repercussions for non-payment, while more advantaged persons can more readily avoid this cycle by making regular payments. Yet, questions remain regarding the influence of various factors that drive this “two-tiered” system of LFO debt persistence and non-payment sanctions.

This paper advances understanding of LFO repayment by identifying the factors associated with the likelihood that persons pay off their entire debt balance. I estimate the relative impact of individual- and institutional-level factors on full payment, including debt balances, payment plans, repayment period length, enforcement actions, and demographic variables. The analysis features data drawn from a sample of persons sentenced to probation and transferred to the Monetary Compliance Unit (MCU, hereafter), a collections arrangement purported to promote repayment in a feasible manner absent active probation supervision and criminal sanctions (i.e., violations or incarceration) as enforcement actions. Results from bivariate (t-tests) and multivariable (logistic regression) analyses indicate that both individual and institutional factors influence the likelihood of full LFO repayment, with especially strong associations for debt balance amounts, monthly payment amounts, and enforcement action receipt (capias warrant). Within the context of the MCU, these findings suggest that some persons are assessed LFOs in excess of what they can repay and will simply be unable to pay down these debts, even when the collections process is comparatively individualized and non-punitive. Further, civil enforcement actions may be a less harmful alternative to criminal enforcement actions in the collections process. After describing the state of knowledge on LFO assessment and collection, I detail the MCU’s structure to contextualize this analysis, then conclude by discussing takeaways within the context of MCU policy and their broader implications for understanding repayment.

1.1. How Much Debt?

This paper uses the term ‘legal financial obligations’ (LFO) to refer to the various financial penalties that can be assessed as a direct result of justice system involvement, each serving different functions in practice. Fines serve punitive goals, as they are often incurred at sentencing as a standalone punishment or a component of punishment. Restitution achieves restoration by collecting funds as remedy to damages to victims of the crime incident. Fees, costs, and/or surcharges can be incurred at various points during processing, adjudication, and punishment to offset system operational expenses. Policies and laws that establish LFOs and set the parameters around their assessment may be instituted through various methods and levels within the legal and criminal justice apparatus (Shannon 2020; Sobol 2016). LFOs are not an innovation in the era of “getting tough” on crime (i.e., Kirk et al. 2020). What is new, however, is the recent increase in the sheer amount of LFOs enacted and assessed throughout the criminal justice system as an additional punishment that often supplements, rather than replaces, incarceration (Beckett and Harris 2011; Menendez et al. 2019).

LFOs are ubiquitously imposed for criminal cases (Katzenstein and Nagrecha 2011) in sizable but variable total amounts. In Washington state, the average criminal case was assessed over $1128 in LFOs (ACLU Washington 2018); this amount was higher—an average of $2500—for persons convicted of felony offenses (Beckett et al. 2008). In Maryland, individuals owed an average of $750 in parole supervision fees (Diller et al. 2009). Another analysis tabulated an average debt of over $1700 owed by individuals under correctional supervision (Pleggenkuhle 2018). Persons charged with felonies in Alabama possess a median LFO debt of just under $4000 (Meredith and Morse 2017). A recent survey of formerly incarcerated persons found that they owed an average of over $13,000 in fines and fees (deVuono-Powell et al. 2015). Link (2019) found that nearly half of a sample of individuals returning from prison had some amount of criminal justice debt—an average of $872—at the point of reentry. LFO debts can accrue substantially for persons repeatedly cycling through the justice system (Harris et al. 2010) and can create a “feedback loop” of continued post-release supervision (Link 2021).

While substantial in amount, comparisons of assessment and collections data indicate that much of these LFO debts go unpaid One analysis estimated an average of $178 million in outstanding LFOs across eleven states (McLean and Thomas 2007). A recent tabulation of court debt at the national level reported over $27 billion in existing debts summed across 25 states (Hammons 2021). Another analysis found $1.9 billion in unpaid LFOs in three states (Florida, New Mexico, and Texas) accrued between 2012 and 2018 (Crowley et al. 2020). As of 2016, there is over $12 billion in outstanding fines and fees in California alone (Legislative Analyst’s Office 2017). These unpaid amounts are a substantial fraction of all that is assessed: for example, Washington state reportedly collects 23.8% of assessed LFOs (ACLU Washington 2018). Because jurisdictions lack sufficiently detailed data that consistently and systematically track LFO assessment and collection (Hammons 2021; Shannon et al. 2020), these staggering numbers are likely undercounts of the true amount of assessed and (un)collected LFOs. The recognition of this gap between assessment and collection raises numerous empirical and policy questions. This paper focuses on one of those questions—specifically, the process of repayment for individuals possessing LFO debt and captured in these aggregate data.

1.2. Factors Affecting Repayment

While numerous factors may affect individuals’ repayment of LFOs, this paper considers the interplay of individual- and institutional-level factors in this process. Ample research documents the importance of both factors for structuring the LFO landscape, but they are often considered as isolated factors.

Starting at the individual level, a primary—if not the primary—factor of importance is individuals’ ability to pay, given the overrepresentation of persons from socioeconomically disadvantaged backgrounds in the system (Harris 2016). Considering attorney type as a proxy for socioeconomic status, a recent analysis of Pennsylvania court records found that persons represented by indigent defense attorneys repaid less of their LFOs than those represented by private counsel, even though they were assessed comparatively lower totals of fines, costs, and restitution (Ward et al. 2020). Some have characterized their LFO debts as “insurmountable” and “unpayable” relative to their personal finances and costs of other necessities such as housing, utilities, and food (Harper et al. 2021; Shannon et al. 2020, p. 276). Interviews with juveniles owing LFOs and their families indicated that non-payment was common and that this debt burden introduced new and substantial challenges to already strapped family finances (Paik and Packard 2019). For some, LFO debts can have spillover effects for family members and supportive others who provide financial support in repayment (deVuono-Powell et al. 2015; Montes et al. 2021). Although increased income during reentry appears to reduce debt balances (Link 2019), difficulty in repayment is still common because many returning from prison tend to work in low-wage jobs that leave little remaining funds for LFO payments after paying for other necessities (Diller et al. 2009).

Although evidence is limited, individuals’ demographics characteristics also appear to matter for LFO debt and repayment. Permissible ranges for fines, fees, and costs may be set by law and/or policy, but extralegal factors—such as race and age—can also impact the amount and type of assessed LFOs (Ruback 2004). Other studies show that Black persons have significantly higher LFO debt (Edwards and Harris 2020; Shannon 2020) and face more difficulty in repayment (Link 2019; Link et al. 2021). Age is also important, as older persons possess more debt and seem to struggle to repay (Link et al. 2021; Ruback and Clark 2011).

At the institutional level, the laws, policies, and dynamics relevant to the assessment and collection of LFOs across jurisdictions contextualize observed debt amounts and repayment enforcement. At assessment, some LFOs are imposed because characteristics of the ‘instant’ (i.e., current) case trigger statutorily mandated fines and/or fees. For example, felony convictions tend to generate higher LFO amounts than misdemeanor cases (e.g., Beckett et al. 2008), and punishment type—such as post-release community supervision—can increase the amount of debt owed during reentry (Link 2019). These “proximally” understood reasons for LFO debt are attributed to established criteria that guide adjudication and sentencing, including case types, sentencing statutes, and discretionary factors (Spencer-Suarez and Martin 2021). Increasingly, assessments may also be influenced by localized reliance on fees and costs revenue as an alternative to tax increases to fund local criminal justice agency operational budgets (Peterson 2012). This possibility may drive observed and substantial between-county variation in assessment amounts for the same offenses (Greenberg et al. 2015) and owed amounts in rural/non-rural probation agencies (Shannon 2020).

Assessment policy is highly relevant for the process of repayment because the amount of local laws and policies that impose LFOs is inversely related to the likelihood of repayment (Ruback and Bergstrom 2006) and because persons assessed higher amounts of LFOs were less likely to repay these debts in full (Gordon and Glaser 1991). At the point of assessment, many jurisdictions do not conduct and consider individuals’ ability to pay as they are supposed to (Crowley et al. 2020). This inflexibility and inability to alleviate burdens—in isolation or in conjunction with individual factors—may leave people with unreasonable debt burdens relative to their means.

Shifting to collection, localized reliance on fee and cost revenue can influence system functioning and decision making around collections (Atkinson 2016; Graham and Makowsky 2021; Sobol 2016). Varied prioritization of this revenue stream ostensibly generates differentially aggressive collection and enforcement contexts and differential understanding of this responsibility within other agency responsibilities (Hyatt et al. 2020). The specific method of collection seems to matter for repayment, as there may be more unpaid debt in jurisdictions that rely on specialized units within probation departments to collect LFOs (Ruback et al. 2004). The process of collections itself may be counterproductive to stable employment because the “procedural pressure points” involved in collections arrangements can simultaneously prevent additional punishment but have a net negative impact by disrupting work schedules (Cadigan and Kirk 2020). Pattillo and Kirk (2021) recently argued that LFO debts and repayment plans engender a “layaway freedom” that confers an additional layer of punishment—money and time—to those struggling to pay off their balances by requiring additional supervision and court hearing obligations during extended periods of repayment.

What is missing from this body of research is an assessment of how the confluence of these individual and institutional factors affect repayment. Neither exists in a vacuum, yet research has only started to explore how individual-level factors affect LFO payment across jurisdiction-specific contexts of assessment and collection. This research is important because the degree to which a jurisdiction relies on this revenue may shape their approach to LFO assessment and collections; in turn, the pursuant policies within this context impact debt amounts and non-payment consequences for justice-involved persons, which can be severe—such as the loss of a driver’s license, supervision violations, or even incarceration (Friedman 2021; Harris et al. 2017; Needham et al. 2020; Piquero and Jennings 2017; Shannon et al. 2020). This study begins to address these questions by considering LFO repayment in a specific collections arrangement—a setting which I now describe.

1.3. Current Study

This study addresses this research question: what are the individual and institutional factors that differentiate individuals who repay all of their LFO debt from those who have outstanding balances? In turn, what do these differences reveal about debt accrual, ability to pay, and the efficacy of non-payment enforcement actions? The present research setting is within probation, a form of community corrections—a term referring to punishment types administered in non-custodial settings (i.e., outside of prisons and in the community). Most people supervised by community corrections in the United States are on probation: in 2019, this population included just under 3.5 million adults (Oudekerk and Kaeble 2021). People on probation have limited freedom: although they are not incarcerated, they are required to follow numerous conditions set by their supervising officer and/or the judge for the entire duration of supervision. Examples of applied conditions include routine reporting to their supervising officer, avoiding crime and/or substance use, maintaining employment, attending therapy and/or treatment services, and paying restitution and other LFOs. If these rules are violated, individuals may receive more restrictive conditions or, in some instances, incarceration. Community corrections present a unique and interesting context for examining LFO repayment because these agencies can bear dual responsibility for LFO assessment and collection. Probation and parole agencies can impose and collect internally-assessed costs and fees associated with the use of supervision conditions, such as monthly supervision fees or electronic monitoring device operation fees. In some jurisdictions, these agencies are also responsible for the downstream collection of externally-assessed fines, costs, and restitution imposed by the courts and other preceding agencies (e.g., Ruhland 2020; Hyatt et al. 2020).

Focused on a single county agency, this study analyzes data drawn from administrative records kept by a collections sub-unit of an adult probation agency—the Monetary Compliance Unit (MCU), described in detail in Link et al. (2021) and summarized here. The MCU is an administrative entity with the sole responsibility of collecting outstanding LFOs outside the bounds of criminal supervision. Most cases transferred from active probation status to the MCU (88%) are persons who have completed all requirements of probation but for full LFO repayment. The MCU straddles civil and criminal boundaries, accepting cases from criminal supervision and using civil enforcement actions—not violations and incarceration—to encourage repayment. This specific collections approach avoids the facilitation of modern debtors’ prisons by blocking pathways to incarceration for non-payment (e.g., Atkinson 2016; Sobol 2016).

The MCU is an interesting context for assessing the impact of individual and institutional factors on LFO repayment because unit policy purports to use an individualized and realistic approach to LFO repayment. At intake, MCU staff consolidate an individual’s existing LFO debt across all court dockets, present and past, into a single sum at intake. From there, staff collaborate with persons to set plans to pay down this debt balance. Agency policy directs staff to conduct individualized assessments of individuals’ income, balances, and monthly expenditures to determine monthly payment amounts to be realistic, attainable, and tailored to each person’s financial circumstances. These tailored monthly payments are thus roughly indicative of individuals’ socioeconomic status: those with higher means can commit to higher payments; those with lower means agree to pay less. These institutional factors coalesce to generate a repayment context that sets individualized and realistic expectations, representing a sharp departure from more rigid and punitive collection approaches such as that of Ferguson, MO (Atkinson 2016). In this setting, the extent to which individuals appear to struggle to repay in will provide insight into the challenges people may face net of repayment and enforcement policy.

2. Data

I analyze a unique dataset created from two sets of administrative records from a single county adult probation agency. The primary data source is administrative records from the MCU that include all cases from the creation of the MCU to the time of data receipt, a period spanning nearly seven years (November 2012 to November 2019). These records are used to track cases, payment plans and amounts, remaining balances, compliance, and enforcement actions. All financial variables reflect total assessed LFOs collapsed across all types (fines, fees, costs, restitution) and, in accordance with MCU policy, collapsed across all criminal dockets with outstanding balances.

MCU data are supplemented with administrative records provided by the county probation agency containing demographic data. These two datasets were merged to create a single file for this analysis containing financial and background information. The merge process successfully matched 88% of MCU cases with demographic data (5811 of 6634). Data that were unable to be matched either had codefendants (with a shared docket number) or a data entry error in merging criteria. Because of the nature of these omissions, it is unlikely that their exclusion introduces bias into the dataset. For example, existing research does not suggest fundamental differences in persons charged individually or with a co-defendant that would bias inferences about individual factors considered in this analysis. Further, policy does not appear to assess or collect LFOs differently for cases with co-defendants. Finally, data entry errors are plausibly random. For purposes of this analysis, I exclude a small number of cases that originated outside of the focal county to maintain a consistent county context because of the localized importance of LFO policy and assessment (i.e., Graham and Makowsky 2021).

2.1. Measures

To better understand observable differences associated with repayment, the focal dependent variable is a binary measure of whether a case action was marked as paid in full. It equals 1 if the MCU administrator indicated that the full LFO balance (calculated at the time of MCU intake) had been repaid, and equals 0 otherwise.

Independent variables span individual and agency-level factors. Individual factors include debt amounts, criminal history, and demographics. Starting balance tabulates the LFO balance at the time of transfer to the MCU. As stated above, this amount is the sum of remaining LFO balance at the conclusion of the probation sentence and, if applicable, outstanding debt from previous dockets (i.e., cases) in Dauphin county courts. This consolidated amount provides insight into persons’ accumulated LFO debt burdens, an important factor to consider in research (e.g., Pleggenkuhle 2018).

Total dockets indicate the number of dockets with debt balances at the time of transfer to the MCU—in other words, all dockets with remaining debt balances that are tabulated into the starting balance. A docket is roughly synonymous with a case: dockets are the official court record for a criminal incident that is inclusive of all charges, sentences, and assessed LFOs associated with the incident. I top-code this measure at 6 dockets in the analytic sample because it is the highest number of dockets among those paid in full. This decision collapses docket totals for 109 cases that had between 7 and 13 dockets with outstanding debt at MCU intake and allows for better comparisons across subgroups. Docket counts provide intuition about an individual’s prior record, prior ability to pay, and accrual of LFO balances.

Monthly payment amount indicates the agreed-upon sum that the individual is expected to pay towards their balance on a monthly basis. Because this amount is intended to reflect what an individual can realistically and consistently pay, it provides a rough but reasonable approximation of their socioeconomic status. Time on MCU tracks the time that an individual has been actively associated with the MCU and its policies, calculated by the elapsed number of days between the recorded start date and the date of data import (1 December 2019). For individuals who have paid in full, this time window is the elapsed time between the start date and the date of the most recent payment that fulfilled all remaining debt.

The final individual-level variables capture demographic characteristics. Non-white is a dummy variable indicating that a person’s race is identified as Black, Asian, Pacific Islander, or Unknown, with white as the reference category. Hispanic measures whether a person has been identified as of Hispanic ethnicity. Male indicates that a person is a male, with female as the reference. Finally, Age counts the respondents’ age (in years) at the time of data import.

For institutional factors, two dummy variables indicate whether an individual received either of the two non-criminal enforcement actions that can be levied by the MCU for those out of compliance with their payment plans. Contempt is an indicator variable marked as one for persons who received notification of a contempt of court hearing; capias is an indicator variable marked as one for persons issued a capias warrant if they fail to appear at their contempt hearing. The process for initiating either of these actions is as follows: persons may be placed on a “contempt of court” list—and receive written notification of this change—if they have gone six months without making a payment and have not responded to three corresponding warning letters. The contempt process stops if the person makes a payment; the person receives a hearing for being in contempt of court if they do not make a payment. The hearing is an opportunity to make a payment or re-negotiate payment plans, but can generate a capias warrant on a discretionary basis that requires attendance at a civil hearing if the person fails to appear. Importantly, neither of these actions have criminal punishment repercussions (i.e., supervision violation or incarceration for non-payment) nor does their issuance transfer individuals back to criminal supervision.

2.2. Analytic Plan

The analysis answers research questions regarding LFO debt and repayment using descriptive and multivariable analyses. After providing sample-level descriptive statistics that indicate the amount of debt, commonality of repayment, and sample demographic information, I examine differences in group means between cases with and without remaining debt balances. I then estimate logistic regressions to ascertain the degree of influence various observed covariates exert on an individual’s likelihood of fully repaying their debt. The analytic sample (n = 5043) is comprised of MCU cases that include full information on the aforementioned covariates. Listwise deletion resulted in the loss of a small amount of cases with missing information on key variables including monthly payment plan amount (10.5% missing; n = 610), male (3.5% missing, n = 203), age (2.1% missing, n = 122), and race (1.74% missing, n = 101). Missing values on these variables result from incomplete administrative tracking of MCU cases within their internal records. All cases with missing data on monthly payment plan amounts have paid their LFO balances in full, spending an average of 0.12 days (range: 0 = 2.86) in association with the MCU. These patterns suggest that cases with missing payment plan information are persons who quickly absolve their LFO balances at the point of intake, perhaps before collaborative meetings with unit staff to set payment plans. While their exclusion is a minor limitation of this study, the sample still includes cases with monthly payment data and comparably short durations of association with the MCU, allowing for the investigation of the dynamics of repayment for similarly situated persons in this analysis.

3. Results

3.1. Describing LFO Debt Balances

Table 1. Sample Descriptive Statistics.

Variable	Mean or %	Median	Minimum	Maximum
Starting Balance	$4582.74	$1826.30	$6.36	$1,210,611.00
Number of Dockets	1.79	1.00	1.00	6.00
Monthly Payment Amount	$55.82	$50.00	$1.00	$15,275.50
Balance Paid in Full	31.8%	--	--	--
Time on MCU (Days)	792.47	287.00	0	2495.00
Contempt Hearing	4.5%	--	--	--
Capias Warrant	2.9%	--	--	--
Non-White	48.44%	--	--	--
Hispanic	8.27%	--	--	--
Male	66.86%	--	--	--
Age	40.74	--	--	--
n	5043

Note: Values computed from analytic sample. Number of dockets is top-coded at 6; time on MCU is in days; MCU stands for Monetary Compliance Unit; Hispanic reference group is non-Hispanic; age is in years.

displays all variables used in this analysis and accompanying descriptive statistics. At the point of transfer to the MCU, cases have an average starting balance of LFO debt of $4582.72, with a median amount of $1826.30. This average debt amount is associated with an average of 1.79 dockets. Considered together, persons transferred to the MCU are responsible for paying this substantial LFO debt that is, for many, a consolidation of nearly two dockets’ worth of outstanding LFOs. It signals that repeated system involvement is common and may drive some debt accrual that, in turn, drives higher debt burdens at the point of intake (and consolidation).

On average, the MCU sample agrees to make monthly payments of $55.82 (median $50.00) towards their starting balance. This amount is purported to be set collaboratively and attainably to the individual’s ability to regularly fulfill this obligation and other monthly expenses: MCU staff are instructed to “aim low” to promote success and compliance (Link et al. 2021). This variable has a substantial range: at the lowest end, two individuals are set to pay $1.00 per month; at the highest end, one individual pays $4000 per month, while another pays $15,275.50. Because of the process used to determine this amount, it appears that the MCU oversees individuals of highly variable socioeconomic means—as gleaned by the amount they are asked to pay monthly—to repay LFO debt.

Since its inception in late 2012, just shy of one-third (31.8%) of cases transferred to the MCU have paid off their balance in full. The average number of days spent in affiliation with the MCU is approximately 792, or just over two years. Some spend zero days on the MCU, indicating that some have an ability to quickly make a single payment at the point of transfer to absolve debt obligations and avoid this form of institutional engagement altogether. Others have spent nearly 7 years (6.8) with the MCU, a duration that covers approximately the entirety of the unit’s existence. Because most (88%) are transferred to the MCU at the ‘max out’ date of the probation sentence, the average MCU case is connected with this county probation agency for nearly two additional years after the completion of their criminal sentence but before the full repayment of LFO debt. While the MCU involves far less intensive requirements and far less severe sanctions than probation, individuals still remain connected to the agency for an extended period beyond the initial probation sentence.

With respect to demographics, just under half (~48%) of individuals transferred to the MCU are non-white, a collapsed category that includes persons identified as Black, Asian, Pacific Islander, Native American, or unknown race. Approximately eight percent of persons are identified as of Hispanic ethnicity, and the average age of MCU-affiliated persons is slightly older than 40 years of age. Finally, institutional factors of enforcement actions for non-compliance with payment plans are relatively uncommon in the MCU. Since its creation, only 4.5% of cases have received a contempt hearing and even fewer (2.9%) have been issued a capias warrant. This infrequent usage suggests that unit staff indeed take all possible action to bring persons back into payment compliance and use these non-criminal sanctions as a last resort.

Table 2. Differences in Means between Persons Who Have and Have Not Fully Repaid Starting LFO Balances.

Covariate	Not Paid in Full	Paid in Full
	Mean (SE)	Mean (SE)
Beginning Balance	$5983.63 (613.78)	$1584.69 *** (106.68)
Total Dockets	2.00 (0.03)	1.50 *** (0.02)
Monthly Payment Amount	$43.69 (1.38)	$81.79 *** (9.82)
Ratio of Monthly Payment to Beginning Balance	2.43% (0.07)	9.33% *** (0.50)
Time on MCU (Days)	913.72 (11.04)	533.21 *** (10.71)
Contempt Hearing	0.04 (0.00)	0.05 (0.01)
Capias Warrant	0.02 (0.00)	0.04 *** (0.01)
Non-White	0.51 (0.01)	0.43 *** (0.01)
Hispanic	0.093 (0.00)	0.07 + (0.01)
Male	0.67 (0.01)	0.66 (0.01)
Age	40.16 (0.21)	41.96 *** (0.32)
n	3437 (68.15%)	1606 (31.85%)

Note: + = p < 0.10; * = p < 0.05; ** = p < 0.01; *** p < 0.001.

assesses differences between cases with and without outstanding debt balances by presenting group-level differences in covariate means (t-tests) for persons who have and have not paid their LFO balances in full. There are many significant differences in individual and institutional factors between these groups. Those who have not paid in full have a significantly and substantially higher starting balance than those who have paid in full—a difference of nearly $4400. While it is intuitive that the amount to be repaid affects the likelihood it is repaid, this difference is remarkable in its sheer magnitude. Relatedly, those who have paid in full have a half docket less associated with their starting balance—in other words, they have fewer dockets with remaining debts that are consolidated at the point of transfer. This difference amounts to a quantitatively smaller burden that may materialize because these persons have shorter prior histories of court involvement or have increased ability to pay down prior debts and current debts during the period of active criminal supervision.

The difference in monthly payment plan amounts is also striking: those who have paid in full also agree to an average monthly payment that is nearly double ($81.79) that of those with remaining balances ($43.69). This difference equates to an increased ability to repay LFO debt at a faster rate. To this point, I calculated monthly payment plan amount as a percentage of starting balance to obtain a sense of individuals’ relative capacity to commit to repay their total LFO debt. Across the entire sample, the monthly payment averages approximately 4.6% of the beginning balance. The monthly payment for those with outstanding balances is approximately 2.5% of their starting balance; those who have paid in full paid committed to pay just over 9% of their balances each month. These values are significantly different and indicate a sizable gap in agreed-upon payments between the two groups. As described above, payment plans are set at amounts that individuals can reasonably pay on a regular basis in conjunction with their income and total expenses. In this way, considering monthly payment as a proxy for socioeconomic status, these results signal stark differences in the advantage levels between these groups. There are also large differences in time with the MCU: those who have fully repaid are associated with the unit for approximately one year less (1.46 years, approximately) than those who still owe (approximately 2.5 years). Accordingly, these persons can more readily make repayments, shortening their time of affiliation with the MCU within this county probation agency.

There is a significantly higher proportion of non-white persons and a marginally significant lower proportion of Hispanic persons in the group with remaining debt balances compared to the group paid in full. Put differently, those who have fully repaid their LFO debts are more white and more Hispanic. While there are no detectable differences in the sex breakdown across groups, those who have paid in full are nearly two years older than those with balances.

With respect to institutional factors such as enforcement actions, there is no detectable difference in the use of contempt hearings between groups with and without balances. However, there is a significantly higher prevalence of the use of capias warrants among persons who have paid in full relative to those who have outstanding balances. This increased proportion indicates more usage of the comparatively severe compliance lever among those who, thus far, appear to face less difficulty in making payments. In results not shown, people who have paid in full and received a capias had an average monthly payment of $60.28, an amount far lower than the group-level average. This proportion suggests a within-group difference where a subset of individuals generally have the capacity to pay but potentially face some challenges in doing so—potentially explaining their receipt of capias warrants and their full LFO debt repayment.

3.2. Multivariable Results

Descriptive differences across groups preliminarily indicate several important differences between those with and without remaining balances: those who have fully repaid balances have a smaller starting burden, commit to higher monthly payments, have fewer prior dockets, and spend less time in affiliation with the MCU. I now present results from logistic regression models that estimate how each covariate impacts the probability of full LFO repayment net of other included individual and institutional factors.

Table 3. Predictors of Full LFO Repayment.

Covariate	Odds Ratio	Standard Error	Marginal Effect	Standard Error
Beginning Balance ^	0.25 ***	(0.01)	−0.20 ***	(0.01)
Total Dockets	1.27 ***	(0.05)	0.03 ***	(0.01)
Monthly Payment Amount ^	3.79 ***	(0.25)	0.19 ***	(0.01)
Time on MCU (Days)	1.00 ***	(0.00)	0.00	(0.00)
Contempt Hearing	1.59 *	(0.29)	0.07	(0.03)
Capias Warrant	3.63 ***	(0.80)	0.20	(0.04)
Non-White	0.67 ***	(0.05)	−0.06 ***	(0.01)
Hispanic	0.70 *	(0.10)	−0.05 **	(0.02)
Male	0.90	(0.07)	−0.02	(0.01)
Age	1.02 ***	(0.00)	0.00 ***	(0.00)
Constant	47.38 ***	(18.88)	--	--
n	5043

Note: Estimates obtained from logistic regression models. ^ = variable has been transformed to its natural log. + = p < 0.10; * = p < 0.05; ** = p < 0.01; *** p < 0.001.

presents results as odds ratios and transformed marginal effects. I focus my presentation and interpretation of results on marginal effects that indicate the average impact of the covariate on the probability of the outcome. Marginal effects are especially advantageous for this discussion because they utilize a common scale across covariates of different metrics, allowing for simpler comparisons in relationship size (Long and Mustillo 2021). Substantively, these values indicate the impact of the regressor on the probability that an individual has repaid their debt balances in full. These models include the natural logarithmic transformation of Beginning Balance and Monthly Payment Plan Amount variables to account for their highly left-skewed distributions.

Results from logistic regression models, presented in Table 3, indicate that most covariates significantly predict the probability of full LFO debt repayment. Starting with reduced odds of full repayment, all factors that correlate with a reduced likelihood are at the individual level. Higher beginning balances are associated with significantly lower odds of LFO repayment. The marginal effect of −0.20 indicates that each 1% increase in beginning balance amount decreases the probability of the likelihood of full LFO repayment by approximately 20%. In other words, a small change in starting balance has a substantial impact on whether that balance will be fully repaid. Time in association with the MCU also has a significant and negative impact (0.999) on the odds of full repayment. However, the size of this impact is substantively small: the marginal effect of −0.0001 means that a 1% increase in the number of days in association with the MCU reduces the likelihood of full repayment by 0.01%. One advantage of the MCU model is that it allows for extended time for LFO repayment without extending the duration of probation supervision and associated reporting, condition compliance, and other requirements. This extended time to repayment, however, is nearly non-impactful in these data, especially relative to other considered factors.

Relative to white persons, non-white persons have significantly lower odds of full LFO repayment. Considered as a marginal effect, non-white persons are 6% less likely than white persons to have paid back their LFO balances. Despite results from t-tests indicating differences in representation across groups, multivariable results indicate that Hispanic persons have a lower probability of being paid in full than non-Hispanic persons. Substantively, Hispanic persons are approximately 5% less likely to fully repay their LFO debt than non-Hispanic persons. Considered together, race and ethnicity correlate with repayment likelihood in a way that reduces the likelihood of debt repayment for nonwhite and non-Hispanic persons.

Shifting towards factors that increase the odds of full LFO repayment, several factors at the individual and institutional levels impact this likelihood. At the individual level, higher monthly payment plan amounts significantly increase one’s odds of full LFO repayment. A 1% increase in monthly payment amount correlates to a nearly 19% increase in the likelihood of full starting balance payment, on average. Similar to findings regarding starting balance, an incremental increase in the amount of money committed to be paid on a monthly basis carries a sizable impact on repayment of the total LFO sum. In other words, small increases to regular payment amounts have a large impact on whether these debts will be fully paid. Of course, the ability to commit to those small increases varies across individuals’ financial status net of other monthly expenses, indicating that those with less means may struggle to pay off their LFO balances.

A higher number of dockets (cases) recorded and consolidated at the time of MCU intake is associated with an increase in the probability of full LFO repayment. Added dockets with LFO debts increase the likelihood of complete repayment of starting balances by approximately 3%, a small but non-trivial amount. There are no significant differences between men and women in the odds of full balance repayment, despite that the sample is predominately comprised (approximately two-thirds) of men. Older age is associated with a significantly higher probability of completely paying off LFO balances: a 1% increase in someone’s age correlates with an increased likelihood of full LFO repayment, but by a small amount of 0.5%.

Institutional factors—specifically, enforcement actions of contempt hearings and capias warrants taken to bring individuals back into compliance with stipulated payment plans—correlate with significantly increased odds of full LFO balance repayment. Marginal effects bring specificity to the magnitude of these positive associations. Receipt of a contempt hearing raises the likelihood of full repayment by approximately 6%, while receipt of a capias warrant raises the likelihood of repayment by 19.8%. These relationships suggest that the use of these civil enforcement actions—especially capias warrants—does correlate with eliciting full repayment. The magnitude of the impact of a capias warrant is the second largest of all considered covariates, falling just behind starting balance and slightly above monthly payment amount. In comparing these relationship sizes, it is apparent that both individual and institutional factors significantly impact the likelihood of full LFO repayment.

4. Discussion

This paper examined the relative influence of individual and institutional factors that predict repayment of outstanding LFO debt in a sample of individuals sentenced to probation and transferred to the MCU, a subdivision of a county adult probation agency focused on collections. MCU policy generates a collections strategy purportedly geared towards encouraging full LFO repayment without the use of harsh penalties through tailored and realistic payment plans paired with the use of civil—not criminal—enforcement mechanisms. Findings reveal that that both individual- and institutional-level factors affect individuals’ repayment of LFO debts, and that some persons struggle with repayment in this context. In this section, I discuss the meaning of three key findings for LFO policy and reform and how they may inform directions for future research.

The first major takeaway is that individual factors matter: relative to those with outstanding balances, those who repay their LFO debts in full have more means and resources to commit to absolving a financially smaller burden. Conversely, those who have not fully repaid their LFO debt have higher balances and commit to smaller payments. This point is magnified when considering payment amounts relative to debt amounts. Those with outstanding balances commit to paying an average of 2.5% of their debt each month, a ratio that translates to over three years (40 months) to full repayment—if all payments are made in full every month. It appears that some will simply struggle to get out from under this burden when assessed amounts are in excess of their personal means. This finding suggests an additive impact of the “proximately” understood LFOs to individual’s socioeconomic status. It adds to understanding of the ways that LFOs are uniquely challenging for the poor (Harris 2016; Slavinski and Spencer-Suarez 2021) by quantifying the gap in amount to repay and means to repay between more and less advantaged persons on probation.

This conclusion is magnified by the institutional context in which these findings are observed. Persons struggle to repay even in a process of collections oriented towards attainability and leniency that does not impose additional fees, fines, or other costs incurred by active status on supervision and/or payment compliance. For some, these characteristics of MCU policy intended as beneficial to individuals’ repayment process simply do not alleviate the burden of LFO debt assessed in amounts that cannot feasibly be repaid. There is limited institutional capacity to assuage these burdens by the time persons are transferred to the MCU’s jurisdiction. Accordingly, these findings contribute to broader discussions regarding the appropriateness of levying LFOs against the poor (Harris 2016). At the very least, they suggest a need to targeting front-end assessments—rather than leniency in post-assessment collections—as an important policy reform (Link et al. 2020). Specifically, there appears to be a need for reforms that systematically implement ability to pay assessments and ensure their results guide decision making at the various points at which LFOs are incurred, with the goal of avoiding debt assessment that is unrealistic relative to one’s means.

A second takeaway is that one key characteristic of the institutional approach to collections adopted by the MCU—allowing additional time to repay—was not substantively important to the probability of repayment. One aspect of the MCU is that it does not require full LFO repayment as a prerequisite to discharge from probation. Instead, the arrangement offers additional time without additional supervision—and associated risks of violation—to pay off LFO balances. Additional days on the MCU were significantly correlated with an increased odds of paying back in full, but this relationship was miniscule (0.01%), especially in comparison with factors such as starting balance and monthly payment amount. This finding dovetails with the first takeaway, as it appears that extended time for repayment will not help those who are unable to repay.

This extra time does not appear to increase repayment, but needs to be further studied to understand how it may be counterproductive to collections and individuals. It effectively connects persons to the justice system for an extended period of time following their probation sentence, albeit to an administrative unit that does not include pathways to incarceration or punishment. This structure does not quite approximate the “shadow carceral state” (e.g., Beckett and Murakawa 2012; Link et al. 2021), but does echo concerns about processes of “layaway freedom” during the repayment process (Pattillo and Kirk 2021). In essence, these individuals are in a liminal space where they do not have probation obligations but remain connected to the system and at some risk of civil enforcement actions. This reality prompts questions about the ramifications of an endless connection to the unit and the system, even to an administrative unit, for those who appear unlikely to ever be able to fully repay their balances. This is important because of research highlighting spillover effects of this tethering during repayment for employment (Cadigan and Kirk 2020). In addition to assessment reform, expanding post-assessment waivers may be important for helping these persons get out from under debts in a ‘reasonable’ time period, however defined.

The third takeaway is that the enforcement actions that are a distinct characteristic of MCU policy were sparsely used and correlated with an increased likelihood of full LFO repayment. The infrequency of their usage suggests that these actions are not aggressively used within a ‘feedback loop’ fueled by self-sustaining interests (e.g., Atkinson 2016)—a notable finding given that law in this state allocates fee revenue towards county probation operating budgets. Their impact on the probability of full LFO repayment—an increase of 6% (contempt) and over 19% (capias)—suggests that these civil enforcement actions can incentivize payment. While this inquiry lies outside the scope of available data, there are (at least) three potential mechanisms that may be driving this latter association. Because MCU policy advises staff to conduct extensive outreach and communication prior to and simultaneous to issuing either of these actions, these efforts may be encapsulated in the legal action itself and drive the positive association. Alternatively, contempt hearings allow for the recalibration and downward adjustment of payment plans, which may allow individuals to more routinely make their monthly payments and repay their balances if initial plans were too high. Finally, descriptive analyses indicated that those who have repaid their LFO balances and received an enforcement action had monthly payment amounts between those who have fully repaid and received no enforcement action and those with outstanding LFO balances, suggesting the existence of a group that generally had the means to repay but were also responsive to enforcement actions. Future research with more granular data on the timing and amounts of payment is needed to identify the mechanisms through which receipt of these civil enforcement actions correlates with a rise in the likelihood of repayment.

These results indicate that non-criminal enforcement does not necessarily result in less effective enforcement, raising questions regarding the relative efficacy and efficiency of civil versus criminal sanctions for non-payment. Criminal sanctions have more obviously counterproductive consequences for repayment, to the extent that enhanced supervisions conditions or incarceration disrupt employment as the major funding stream for repayment. The consequences of these civil enforcement actions are less clear. While they would seem to avoid the worst harms associated with criminal sanctions, at the very least, the amount of officer communication, home visits, and/or court hearings entailed in either can be cumbersome and disruptive (e.g., Cadigan and Kirk 2020; Pattillo and Kirk 2021).

These quantitative analyses indicate the combined influence of individual and institutional factors on the likelihood of LFO repaymen, and provide a foundation for future inquiries. There is a specific need for qualitative research that builds on this knowledge base by capturing the lived experiences of having and repaying LFO debt for persons affiliated with the MCU and other institutional collections arrangements. Inquiries might investigate the degree to which the “aim low” approach is understood as individualized or how the MCU approach is understood as a distinct and an intended improvement from keeping individuals on typical probation supervision during repayment periods (e.g., Pattillo and Kirk 2021). Further, findings highlight the need for broader efforts to start to define the landscape of collections policy. The current lack of systematic documentation of the various LFO collections arrangements and policies used across jurisdictions makes it challenging to assess the degree to which MCU collections policy is idiosyncratic or representative of other collections policies and prevents drawing any conclusions about the generalizability of findings. Given the observed importance of enforcement actions net of monthly payments and debt balances, it is reasonable to expect that institutional factors will be important for repayment in different contexts—and perhaps even more important in arrangements using a more punitive enforcement and sanctioning approach.

Limitations

These novel data provide new insight into how people fare with LFO repayment in a specific policy context, but have some limitations that preclude additional analyses across several key dimensions. The data are cross-sectional, providing a snapshot of repayment over the MCU’s existence until the point of receipt. They are not longitudinal nor sufficiently granular in nature to pursue more a dynamic exploration of LFO repayment, including the regularity and/or pacing of payments across landmark events (i.e., intake, enforcement action, and case closure) and the exact accrual of LFOs—parsed out across fines, fees, and restitution—across dockets compiled at intake. Balance data are limited to the amounts owed at the time of intake for those transferred to the MCU, preventing comparisons of MCU balances to total assessed balances and to persons on remaining on active supervision. More direct measures of socioeconomic status, such as income, employment status, or attorney type, would have allowed for more direct assessments of feasibility of repayment. Finally, the analysis could have been enhanced with information on the source of payments—for example, from income or supportive others.

5. Conclusions

This analysis revealed the importance of individual and institutional factors for LFO repayment. Persons who pay their balances in their entirety are more advantaged in numerous ways than those who do not. People who face smaller debts and can pay more on a routine basis are more likely to pay off their LFO debt, while others seem to lack the means to make sufficiently sized payments to chip away at their LFO balances. The significance of civil enforcement indicates a need for exploring this alternative to criminal sanctions for non-payment. Overall, findings observed within the context of this specialized collections unit push reform attention to the point of assessment and highlight a need for methods to better calibrate LFO assessment amounts to realistic sums to avoid imposing LFO debts that are simply unpayable relative to one’s means.