Bankruptcy in Black and White: The Effect of Race and Bankruptcy Code Exemptions on Wealth

Matthew Bruckner,Raphaël Charron-Chénier,Jevay Grooms January, 2023

Abstract

Bankruptcy law in the United States is race-neutral on its face but, in practice, race matters in bankruptcy outcomes. Our original research provides an empirical look at how the facially neutral laws that allow debtors to retain assets in bankruptcy cases result in disparate outcomes for Black and white debtors. Racial differences in asset retention in bankruptcy cases play a role in perpetuating wealth inequality between Black and white debtors.

Existing bankruptcy data lacks individual-level characteristics such as race, which inhibits researchers’ ability to adequately assess biases or unintended consequences of laws and policies on subsets of the population. Thus, we construct a novel data set using bankruptcy data from Washington D.C. in 2011 and imputing race. The data demonstrates that facially race-neutral bankruptcy laws contribute to racially disparate outcomes by allowing white debtors to keep larger amounts of both personal and real property.

First, exemption laws allow every bankruptcy filer to retain some personal property even if they do not repay their creditors in full. At the median, white filers in the District of Columbia claimed $10,150 in exemptions, relative to $8,359 for Black filers. In other words, the median white filer kept roughly $1,800 more of their property than Black filers, despite reporting similar overall personal property values.

Second, exemption laws allow every bankruptcy filer to retain some (or all) equity in their home. Unlike personal property, where Black and white debtors enter bankruptcy with about the same amount of property, white debtors enter bankruptcy with more home equity than Black debtors ($585,000 compared with $251,600 at the median). Unsurprisingly, then, white debtors also leave bankruptcy with more home equity (e.g., the median Black filer retains roughly 80% less in home equity than white filers).

Although bankruptcy laws do not inflate the value of white filers’ personal or real property values relative to Black debtors, our exemption rules contribute to white debtors leaving bankruptcy with greater wealth than Black debtors. By protecting certain assets like home equity, which are unevenly distributed in our sample across Black and white debtors, bankruptcy law appears to play a role in perpetuating wealth inequality. Even where assets are more evenly distributed, as personal property was in our sample, bankruptcy law leaves Black debtors with a less robust “fresh start” than white debtors.

2. Personal Property Exemptions Claimed on

Introduction

In the United States, debtor-creditor laws produce racially disparate outcomes despite being facially race neutral. Exemption laws are no different. This Article is among the first to examine the racially disparate results of applying race-neutral exemption laws in bankruptcy, results anticipated by Professor Mechele Dickerson in two ground-breaking articles on how race matters to bankruptcy law.¹

Among other results, we find that exemption laws advantage white debtors relative to Black² debtors by allowing homeowners and holders of certain personal property to retain some or all their equity in those assets when they exit bankruptcy. Some of these racial differences might be explained by higher asset values at the time white and Black debtors enter bankruptcy. For example, white debtors tend to have more equity in their homes when they enter bankruptcy and so they leave bankruptcy with substantially more home equity than Black debtors. However, white debtors in our sample retain a larger percentage of their motor vehicles’ value than Black debtors despite reporting lower values of this type of personal property when they filed bankruptcy. In these ways, exemption laws appear to aid white debtors over Black debtors in retaining wealth post-bankruptcy.³

In her earlier work, Professor Mechele Dickerson argued that bankruptcy law favors certain white debtors over debtors from other racial groups by, among other things, allowing debtors to retain their equity in assets that white debtors are more likely to own.⁴ This Article provides some of the first empirical evidence for Professor Dickerson’s theory by examining the role that exemptions (the method for providing preferential treatment to certain assets in bankruptcy) play in perpetuating or increasing wealth inequality across Black and white Americans. This Article details the frequency and value of real and personal property ownership in bankruptcy for residents of Washington, D.C. in 2011, including homes, vehicles, retirement accounts, apparel, jewelry, cash and cash equivalents, and other household goods. The first two—houses and cars—turn out to be the most important.

This Article offers two primary contributions. First, we created a new dataset based on a random sample of bankruptcy petitions drawn from the United States Bankruptcy Court for the District of Columbia, all of which were filed in 2011. This dataset will be made available to other researchers in the future.⁵ Second, we describe and quantify racial disparities in bankruptcy cases that exist despite our facially neutral exemption laws.

The remainder of the Article proceeds as follows. In Section I, we provide some background, both on the racial wealth gap and on the relevant existing literature. In Section II, we provide an overview of consumer bankruptcy law with a focus on the distinctions between chapter 7 and chapter 13. In Section III, we provide our data analysis, including a separate analysis for real and personal property. The most important personal property categories are vehicles and retirement accounts. A conclusion follows. Additional tables are available in Appendix A, and Appendix B provides additional details about our dataset, including its limitations and how we estimated the racial identity of debtors in our sample.

I. Background

A. The Racial Wealth Gap

This Article explores wealth differences between Black and white Washingtonians who filed for bankruptcy in the District of Columbia in 2011. Before doing so, however, we note the pre-existing disparities in Black and white wealth in America, generally. The racial wealth gap has always existed in America. According to data from the Survey of Consumer Finances, the median white household in America has a net household wealth that is more than eight times greater than the median Black household.⁶ This stark disparity is true across the spectrum of income distribution in the United States. For example, one study found that more than 96% of the top 1% of the nation’s wealthiest Americans were white and just under 2% Black.⁷ Conversely, this same study found that white households near poverty hold $18,000 in wealth compared to zero dollars in wealth for comparable Black households.⁸

In Washington, D.C., the wealth gap between Black and white households was greater than the national wealth divide. According to a report from the Urban Institute, white households in the District of Columbia had a net worth more than eighty times greater than the typical Black household.⁹ Put differently, the racial wealth gap was nearly ten times larger in the District of Columbia than exists nationally. The Urban Institute reported on the value of various categories of personal property, with white families being more likely than Black families to have vehicles,¹⁰ retirement funds,¹¹ liquid assets,¹² checking accounts,¹³ and other financial assets, such as stocks, bonds and mutual funds.¹⁴ In the aggregate, the median value of liquid assets owned by U.S. Black households was only $5,000 and merely $2,100 for African Black households.¹⁵ By contrast, white “households had the highest median value of liquid assets,” at $65,000.¹⁶

The Urban Institute authors also focused on homeownership rates and home values. They reported that Washingtonians—both Black and white—own homes at higher rates than people in other regions.¹⁷ Although white Washingtonians are more likely to own homes than Black Washingtonians, the racial homeownership gap is smaller in the District than it is nationally.¹⁸ This is important because home ownership has traditionally been an important path to wealth building in this country and has, therefore, been an important component of the racial wealth gap.¹⁹ Unfortunately, the Urban Institute also reports that the homes of Black families in the District are valued “significantly lower” than those owned by white families.²⁰

B. Literature Review

Unfortunately, administrative data does not capture racial and ethnic dynamics of bankruptcy. The most important source of this data is the Consumer Bankruptcy Project (“the CBP”).²¹ As a result, the empirical literature on bankruptcy law and race is limited.²² Thus, the literature on bankruptcy, race, and wealth is even less prevalent.²³ Still, our Article intersects with three strands of extant literature.

First, this Article adds to the empirical literature assessing whether bankruptcy filers are abusing the consumer bankruptcy system. There is a long-standing concern that some people engage in strategic, financial calculations when filing bankruptcy instead of filing bankruptcy because of true financial need;²⁴ that a large volume of bankruptcy filings in times of prosperity are explainable only if people are using bankruptcy as “a means of violating promises willy-nilly.”²⁵ Creditors, particularly the major credit card lenders, pushed this narrative when lobbying for major changes to the bankruptcy code that were eventually enacted in 2005. They asserted that many “bankruptcy filers were opportunists who borrowed without intending to repay and harmed other borrowers by filing when they had not experienced an adverse shock.”²⁶ And Congress took heed.²⁷ The 2005 revisions to the Bankruptcy Code—one of the most significant overhauls of the bankruptcy system—were even called the Bankruptcy Abuse Prevention and Consumer Protection Act (“BAPCPA”).²⁸ Even the order of the words in the Act’s title demonstrates that Congress’ primary concern was preventing bankruptcy abuse.²⁹

In this Article, we contribute to this debate by examining debtors’ assets rather than their income and expenses. It is difficult to obtain data on the assets of individual debtors (though aggregate data is available) and, therefore, this approach is fairly uncommon.³⁰ As one research team noted, “assets and asset exemptions received virtually no attention in the bankruptcy debates.”³¹ That said, some other scholars have previously used homeownership as a proxy for wealth generally and have used this data to study bankruptcy filing patterns.³² By contrast, our study allows us to study debtors’ wealth directly and, therefore, adds to the literature.³³

Second, our Article also adds to the “thin” literature on automobile ownership and financial distress.³⁴ In Driven to Bankruptcy , the authors “describe the relationship between automobile ownership, financial distress, and bankruptcy” and they “identify a subset of debtors, constituting about a third of bankruptcy filers, who come to bankruptcy owning automobiles and little else.”³⁵ These automobile bankruptcies are more likely to be filed by African-American debtors.³⁶ The authors found that African-American debtors are less likely to own cars (82.1%) than non-African-American households in bankruptcy (89.1%), but that African-American debtors are more likely to own cars (55.1%) if they don’t also own a house (53.8%).³⁷

Despite high rates of car ownership, the authors reported that slightly less than a quarter (22.4%) of debtors claim an automobile exemption.³⁸ In terms of value, the median chapter 7 car-owning filer’s “most valuable automobile . . . is worth approximately $5400, which is roughly in the same price range as the trade-in values for a typical five-year-old Ford Focus or a six-year-old Toyota Corolla.”³⁹ The median chapter 13 carowning filer’s “most valuable car is worth $8,928, which is closer in value to a two-year-old Ford Focus or three-year-old Toyota Corolla.”⁴⁰ As a percentage of their total assets, “the median household that filed bankruptcy in our sample entered bankruptcy with cars accounting for about 20 percent of their total assets, and the average household came to bankruptcy court with cars as one-third of their total assets.”⁴¹ When broken down by chapter choice, “despite having less valuable cars, chapter 7 filers have more of their wealth concentrated in cars than chapter 13 filers (23.2 percent versus 13.8 percent).”⁴² Additionally, “[t]he median car owner who claimed an exemption had $2,932 of equity in the claimed car that was exempt from the bankruptcy estate. Chapter 13 filers are more likely to have nonexempt car equity they would lose in a chapter 7 filing (26.1 percent versus 20.4 percent); chapter 13 filers also have higher average amounts of nonexempt car equity ($1268 versus $720). (The medians in both chapters are zero.).”⁴³

Unusually, the authors of Driven to Bankruptcy also examined the debtors’ race and how it intersected with their findings. The authors reported that “African American households enter bankruptcy with slightly more valuable cars at the median,”⁴⁴ And that “[a]lthough African American households are less likely to own cars than other households, this reverses when homeownership is removed from the analysis, and African American households are more likely to own a car if they do not own a home.”⁴⁵ As reported below, our results sometimes align but sometimes differ from theirs.⁴⁶

Third, our work intersects with the limited research on race and bankruptcy. That work has been limited because it is difficult to test some theories about bankruptcy filers empirically. Individual-level bankruptcy data of any sort are not readily available to analyze.⁴⁷ Researchers without access to data from the Consumer Bankruptcy Project have needed to rely on creative empirical strategies, including surveys of bankruptcy professionals, online experiments, administrative data from specific metropolitan areas, and court records.⁴⁸ For this reason, an important goal of this project is to help address this problem by creating a unique dataset with which these empirical questions regarding race and bankruptcy can be examined.

In 2018, Professors Robert Lawless and Angela Littwin reported on existing race-related bankruptcy research in Local Legal Culture from R2D2 to Big Data .⁴⁹ In Local Legal Culture , the authors describe the history of race-related bankruptcy research, which they suggest first began to appear in the early 1990s. They focus on the overrepresentation of Black (and sometimes Latino) debtors in chapter 13 cases, cases “that take[] more time, cost[] more money, and [have] a significantly lower discharge rate. In subsequent work with Professors Braucher and Cohen, Professor Lawless and co-authors reported that Black debtors are likely to fare worse than similarly situated white debtors because chapter 13 cases are more likely than chapter 7 cases to end without the debtor receiving a discharge.⁵⁰

Professor Pamela Foohey’s work found that Black churches—those “[c]hurches with predominately Black membership”— appear “in Chapter 11 [bankruptcy] more than three times as often as they appear among churches across the country.”⁵¹ And Professor A. Mechele Dickerson’s work, which drew on summary statistical data (rather than bankruptcy petitions), argued that the bankruptcy system is more likely to benefit white debtors than Black debtors.⁵² Finally, one of the most interesting research projects to date in this area is from Professors Pamela Foohey, Robert M. Lawless, and Deborah Thorne. Their article Portraits of Bankruptcy Filers ⁵³ found that “Black households file bankruptcy at more than twice the rate they appear in the general population.”⁵⁴

In this strand of literature, research has generally shown that Black debtors and entities affiliated with Black people (e.g., Black-owned businesses) fare worse in the bankruptcy process than their white counterparts.⁵⁵ Our contribution to the literature is to look at differences by race in bankruptcy exemptions, which has only been theorized but never empirically examined. And, as reported below, we also discuss how Black debtors appear to fare worse in bankruptcy than their white peers. Before we report our findings, however, it may be useful to introduce consumer bankruptcy. Readers that are already familiar with consumer bankruptcy may wish to skip to Section III.

II. Exemption Law and Consumer Bankruptcy Basics

“From 2009 to 2018 alone, over 11 million individuals filed bankruptcy.”⁵⁶ And, in 2021, there were approximately 400,000 new bankruptcy filings.⁵⁷ Nevertheless, some scholars have suggested that many people that would likely benefit from bankruptcy do not file for various reasons.⁵⁸ Unfortunately, the benefits of bankruptcy are not evenly distributed among racial groups.⁵⁹ To understand why, we briefly explain some bankruptcy basics.

There are two primary options for individuals who seek debt relief through bankruptcy: chapter 7 and chapter 13.⁶⁰ Historically, chapter 7 is favored by approximately two-thirds of consumer debtors.⁶¹ In chapter 7 cases, the debtor keeps their future income but turns over certain current assets to their creditors.⁶² By contrast, in chapter 13 cases, debtors may keep most of their current assets but they must use certain future income to repay their creditors.⁶³ As a practical matter, however, there may be little difference in creditor recovery rates if a chapter 13 debtor lacks any “disposable income” or a chapter 7 debtor lacks any non-exempt property.⁶⁴ Whichever chapter a person chooses, however, a debtor usually starts their bankruptcy case by filing a petition with the bankruptcy court.⁶⁵

Filing a bankruptcy petition creates the so-called “bankruptcy estate,” which comprises all the debtor’s assets at the time of filing.⁶⁶ The estate’s assets—all of the debtors’ houses, vehicles, retirement accounts, jewelry, clothing, guns, pets, furniture, and virtually everything else—may be used to satisfy a debtor’s creditors.⁶⁷ Debtors must disclose these assets on a form called a schedule.⁶⁸ Where there are substantial assets, these assets are often sold, and the proceeds are distributed to creditors of a chapter 7 debtor in partial satisfaction of the debts owed to the creditors. The remainder of the debtor’s debts are then discharged and need not be repaid.

However, whether they file under chapter 7 or chapter 13, debtors may keep some amount of their property, referred to as exempt property.⁶⁹ In cases where debtors do not own substantial assets, this might result in no recovery for creditors, even in a chapter 7 case where creditors are supposed to be repaid from the debtor’s current assets instead of future income.⁷⁰

Exemption laws limit the amount of property a debtor may retain on a category-by-category basis.⁷¹ In the District of Columbia (like in many states), debtors may choose between federal exemptions and the District exemptions, depending on which are more favorable to the debtor.⁷² In the District of Columbia, the choice often depends on how much equity a debtor has in their home. This is because the District exemptions allow debtors to exempt all the equity in their homes,⁷³ but federal law currently allows a debtor to retain only approximately $25,000 in home equity.⁷⁴

In exchange for giving their nonexempt property, if any, to creditors, the remainder of a debtor’s unsecured⁷⁵ debts are discharged (i.e., forgiven) and need not be repaid.⁷⁶ As a result, “[c]hapter 7 is considered to be most useful to people with few assets of value that they want to keep, with considerable unsecured debts that they hope to discharge, and without home ownership.”⁷⁷

By contrast, chapter 13 debtors must give creditors their future “disposable income” over a three- to five-year period.⁷⁸ Only after they have completed their plan payments do chapter 13 debtors receive a discharge of their remaining debts. “In return, the debtor keeps all property, regardless of exemptions, and is allowed to use the property during the pendency of the case.”⁷⁹ Many debtors who have a house or car they want to keep will use chapter 13.⁸⁰ Debtors can pay their attorneys over time in a chapter 13 case, but they generally have to pay their chapter 7 attorneys in full before the case begins.⁸¹

Unfortunately, this narrative about chapter choice being simply about which deal is preferable to the debtor is complicated by the fact that far more chapter 7 cases result in a discharge than do chapter 13 cases. Only about one-third of chapter 13 cases end with the debtor receiving a discharge of their remaining debts.⁸² For the other two-thirds of cases, debtors do not receive the anticipated debt forgiveness and remain “personally liable for all debts not paid during the chapter 13 case, potentially including additional accrued interest and fees.”⁸³

III. Data Analysis

As discussed further in Appendix B, the Article relies on a sample of 501 bankruptcy petitions drawn from the 959 petitions filed in Washington, D.C. in 2011.⁸⁴ For analyses, we narrowed our focus to those debtors who claimed consumer debts only. For this reason, we dropped 48 petitions that reported business-related debt.⁸⁵ This reduced our sample to 453 observations. For analyses that involved the debtor’s race, we only retained petitions for which we could impute race with an at least 50% probability. This required us to exclude 13 additional petitions. We also excluded 54 petitions that were identified as Hispanic, Asian and Pacific Islander, or American Indian and Alaska Native. Thus, for analyses that rely on race, our final analytic sample includes 386 debtors, of which 326 are estimated to be Black, and 60 are estimated to be white.⁸⁶

Research indicates that throughout the U.S., the rate of bankruptcy filings is higher for Black individuals than white.⁸⁷ For example, in 2017 ProPublica asserted that consumer bankruptcy filings were more frequent in majority Black zip codes compared to majority white zip codes.⁸⁸ This was further echoed by Jonathan Fisher’s 2019 research using data from 1994-2009, which found that bankruptcy filers in the U.S. were more likely to be Black.⁸⁹ This holds true in our sample as well. Roughly 72% of debtors in our full sample are Black, a proportion that is less than one and a half times as large as the Black share of D.C. residents.⁹⁰ Although this is lower than some other researchers have found,⁹¹ the difference may be attributable to the large share of Black residents in D.C. relative to other cities.

To understand the role bankruptcy may play in generating or maintaining racial disparities in wealth, we provide four sets of descriptive analyses. First, we compare Black and white debtors across select variables to understand how the Black and white petitioners differ on relevant dimensions (Section III.A).⁹² Next, given the importance of housing wealth as a component of overall household wealth, we examine racial differences in homestead exemptions (III.B). We then provide descriptive analyses of the personal property Black and white households claim to own on Schedule B of their bankruptcy petitions (III.C.1). We end with descriptive analyses of claimed exemptions for those assets, as reported on Schedule C (III.C.2). Results of relevant statistical significance tests are reported in Table A3 of Appendix A.

A. Descriptive Statistics

This Section provides descriptive statistics for key variables comparing white and Black bankruptcy filers in our data. For most variables, we report both mean and median values. Mean values allow straightforward tests of statistical significance, while median values provide useful estimates of central tendency for skewed distributions (as is the case for most of the variables reported).⁹³

Black debtor profiles are similar to those of white debtors across several key variables in our sample. 85% of Black filers report consulting with an attorney compared to 90% of white filers,⁹⁴ and 13% of white filers reported filing jointly with a partner or spouse relative to 10% of Black filers.⁹⁵ 10% of Black filers report a prior bankruptcy case compared to 15% of white filers.⁹⁶ T-tests comparing proportions across Black and white filers indicate that none of these differences are statistically significant.

Debtors in our sample report a monthly mean income of approximately $3,316 (median = $2,852).⁹⁷ Black debtors report lower monthly incomes (mean = $3,223, median = $2,820) than white debtors (mean = $4,043, median = $3,069).⁹⁸ As reported in Table A3 of Appendix A, the difference in the mean incomes of Black and white debtors is not statistically significant ( p = 0.12).

Debtors are instructed to report average monthly income for the calendar year when reporting monthly income. On an annual basis, the income for all D.C. filers is roughly $39,788 at the mean ($34,228 at the median).⁹⁹ In 2011, this would put the petitioners in our sample above the federal poverty line ($22,350 for a household of four)¹⁰⁰ but substantially below the median household income ($63,124).¹⁰¹

Table 2 provides additional information on annualized income for Black and white petitioners. As estimates from the table suggest, at least three-quarters of bankruptcy filers of either race in our sample have an income that falls below the median income for D.C. overall. For Black petitioners, as many as 90% of filers report an income below the D.C. median ($63,124). To provide a comparison, median annualized rent in D.C. that year was $15,576 ($1,313/month).¹⁰² In other words, the median debtor in our sample would spend roughly 45% of their annual salary in rent (assuming they pay the median rent). Overall, these estimates indicate that the median debtor in our sample tends to be poverty-adjacent, meaning their income is closer to the poverty line than to the median income for the area.¹⁰³

Table 2. Selected income percentiles among debtors Minimum 25 50 75 90th Maximum

(median) Overall $1,476 $23,745 $34,228 $48,168 $69,224 $240,000 Black $1,476 $24,001 $33,834 $46,521 $65,855 $240,000 White $2,400 $27,516 $36,828 $56,311 $76,163 $234,295

Black and white filers also differ along other measures of socioeconomic status. A statistically significant difference is observed in mean reported monthly expenditures ( p < 0.05) among Black and white filers, with Black filers spending roughly $1,000 less than white filers at the mean ($376 less at the median) (illustrated in Figures 2a and 2b). Although we observe a substantial difference in the mean reported assets of Black and white filers ($170,381 vs. $225,405), the difference is not statistically significant ( p -value = 0.24). Lack of significance despite the large difference in mean values is likely due to the large variances in reported asset levels caused by the skew in the debt distribution and the relatively small number of white filers in our sample (see table 3).

As shown in Figure 2a, white filers report $4,561 in mean expenditures monthly (median = $3,486), compared with $3,554 for Black filers (median = $3,110). This mean difference amounts to roughly $12,150 annually. For white filers, the mean ratio of monthly expenditures to monthly income is 2.42 (1.03 at the median). The mean expenditure-toincome ratio for Black filers is lower at 1.26 (1.06 at the median). In other words, the average bankruptcy filer’s financial condition is highly precarious, with expenditures greater than income. Even relatively small, unexpected expenditures could overwhelm available resources.¹⁰⁴

As shown in Figure 2b, white filers report $385,555 in debt at the mean (median = $133,392), compared with $251,427 for Black filers (median = $118,502).¹⁰⁵ For both groups, these debt levels are substantially higher than reported assets. White filers report a mean of $225,405 in assets (median = $24,854), compared with $170,366 for Black filers (median = $30,326). The considerable difference across mean and median debts and assets for both groups indicate skew in the distribution of these quantities across debtors.

Figure 2a Figure 2b

Mean Income and expenditure by Mean Asset and Debt by Race

Race

Table 3. Selected asset and debt percentiles among debtors Minimum25 50 75 90th Maximum

(median) Debt

Overall $377 $46,407 $133,382 $369,859 $619,539 $3,751,656 Black $377 $42,800 $118,389 $338,337 $578,213 $3,459,812 White $16,891 $77,049 $133,392 $483,053 $908,867 $3,751,656 Assets

Overall $0 $7,301 $29,586 $307,639 $452,910 $1,934,528 Black $0 $7,293 $30,767 $290,654 $430,209 $1,934,528 White $595 $7,145 $24,854 $372,685 $779,969 $1,157,211

In our sample, roughly 84% of all personal bankruptcy filers opted for chapter 7, with the rest opting for chapter 13. This is similar to trends for all D.C. bankruptcy filings in 2011, in which 83% of the 927 personal bankruptcy petitions were chapter 7 filings.¹⁰⁶ But it is a higher percentage of chapter 7 cases than historic trends in other jurisdictions.¹⁰⁷

We observe a small (and non-statistically significant) difference in the proportion of Black and white debtors who opted for a chapter 13 bankruptcy (18% for white filers vs. 16% for Black filers).¹⁰⁸ Descriptively, this is inconsistent with the findings of previous researchers who found that Black filers are “much more likely” to file chapter 13 relative to white filers.¹⁰⁹

B. Analyzing Homestead Exemptions

In this Section, we analyze data from debtors’ Schedules A and C. Schedule A provides information on the value of debtors’ “homestead”: real property, land and personal property used as a residence (e.g., an RV), and burial plots held at the time of filing.¹¹⁰ Schedule C provides information on filers’ claimed homestead exemptions—the value of this property they aim to keep after bankruptcy. Homeownership plays a significant role in wealth building. The U.S. Census Bureau estimates that the median wealth of homeowners is nearly ninety times greater than nonhomeowners’.¹¹¹ Subsection IV.B.1 details home values identified on filers’ Schedule A. Although we do not explicitly report outstanding debts on the residence, subsection IV.B.2 details the value (equity) of debtors’ homestead exemption claims.

Fewer than half of all debtors in our sample report owning real property (usually homes) in their Schedule A filings. Notably, 48% of Black debtors report owning any homestead property but only 32% of white debtors do so.¹¹² Yet on average, Black filers who own homestead property report lower property values than white filers (42% lower at the mean, 47% lower at the median).¹¹³ White debtors report homesteads valued at $589,174 at the mean (median $585,000), whereas Black debtors report homesteads valued at $286,653 (median $251,600).¹¹⁴ The difference across these means is statistically significant ( p < 0.05), as reported in Table A3 of Appendix A.

Of the debtors who report a homestead on Schedule A, roughly the same proportion of Black and white filers claim some portion of it as an exempt asset on their Schedule C (62% of Black and 68% of white debtors). These exemptions typically represent the value of home equity that filers seek to retain after bankruptcy. The size of the exemption claimed, however, differs substantially across Black and white filers. Among those who claim exemptions, white debtors claim exemptions valued at $285,451 at the mean (median = $259,406), whereas Black debtors claim mean homestead exemptions valued at $109,469 (median = $21,625). Overall, white filers are expected to leave bankruptcy with homestead property valued at roughly two and half times the value retained by Black households at the mean (and at roughly twelve times at the median). The difference across these means is statistically significant under a one-tail test ( p = 0.036) but not a two-tail test, as reported in Table A3 of Appendix A. Put differently, at the mean, Black filers do not retain $217,308 in homestead value in their bankruptcy cases (median = $195,350) while white filers do not retain $393,865 (median = $325,594).

In summary, on a nominal basis, white filers have more non-exempt homestead assets. In part, this is likely because Black filers own less expensive homes than white filers when they file bankruptcy and borrow more heavily to purchase those homes in the first instance.¹¹⁵ Notably, Black filers retain a smaller percentage of their value when they exit bankruptcy. The gap between the property values reported as of the petition date by Black and white debtors and the assets these debtors seek to retain vis-à-vis their homestead exemption grows by roughly ten percentage points at the mean (and roughly 34 percentage points at the median).

1. Homestead Values as Reported on Schedule A

In the District of Columbia, debtors may choose to use the generally applicable federal exemptions or the District-specific exemptions. These are quite similar in most respects. The most significant difference is that, in 2011, the federal exemptions limited debtors to retaining a maximum of $20,000 in their homestead and the District had no maximum homestead exemption.¹¹⁶

As reported in the previous Section, white debtors enter bankruptcy reporting higher homestead values on average.¹¹⁷ This remains true across type of exemption claimed.¹¹⁸ Among filers reporting any homestead assets, white filers who chose the federal exemptions report homesteads valued $208,442 greater at the mean than Black filers, a 77% difference. At the median, the estimates are $328,850 and 130% respectively. For filers opting for the D.C. exemption, white debtors report homesteads valued at $358,742 greater (116% more) than Black debtors at the mean. Equivalent estimates at the median are $429,170 and 163% respectively.

2. Homestead Exemptions Claimed on Schedule C

Given D.C.’s unlimited¹¹⁹ homestead exemption, we expect that debtors claiming the D.C. exemptions would claim larger homestead exemptions than debtors claiming the federal exemption. Indeed, at the mean, Black filers who choose the D.C. exemption claimed a homestead exemption valued 2.1 times greater than the homestead of Black filers who choose federal exemptions (9.5 times greater at the median). In comparison, whites filing D.C. exemptions claim 3.1 times the homestead value of whites filing federal exemptions (8.0 times at the median). The difference between the mean exemption claimed for white and Black debtors grows from $53,166 for those claiming the federal exemption (median = $33,713) to $240,114 for those claiming the D.C. exemption (median = $251,400).¹²⁰

Note that in some instances the claimed homestead exemption may exceed the maximum permissible exemption. To ensure we are accurately describing the data collected, which can include filing errors, we do not alter the reported values or set a maximum.¹²¹

Overall, then, these estimates indicate the following:

1. A greater proportion of Black debtors report owning a

home than white debtors (48% compared to 32%) in our sample, but median homestead values among homeowners are considerably higher for whites ($585,000) than Blacks ($251,600).

2. Similar proportions of Black and white homeowners

claim a homestead exemption (roughly two-thirds), but the median claimed exemption value for white debtors is roughly twelve times larger than the median value claimed by Black debtors ($259,406 vs. $21,625).

3. As a result, Black debtors are expected to leave bankruptcy

with homesteads values roughly 10% of those of white debtors at the median (38% at the mean), even though Black debtors entered bankruptcy with homesteads valued at roughly 42% that of white debtors at the median (48% at the mean).

C. Analyzing Schedule B’s Personal Property Exemptions

1. Personal Property Values as Reported on Schedule B

In addition to homesteads, debtors must identify any personal property they own when filing for bankruptcy.¹²² This is reported on Schedule B and includes (but is not limited to) jewelry, clothing, cash, retirement accounts, and livestock.¹²³ Although not all petitioners owned a homestead, all petitioners in our sample claimed at least some personal property. At the median, Black filers report amounts of (non-homestead) personal property similar to white filers: $10,516 for Black filers compared to $10,367 for white filers.¹²⁴

This similarity in reported personal property is striking because most research suggests that white families typically own a wider and more highly valued array of personal property than Black families.¹²⁵ This may suggest that the distribution of personal property among Black and white D.C. households is more equal than the national average. Alternatively, it could also potentially suggest that personal property ownership is less useful for avoiding bankruptcy for Black households than for white households in D.C., resulting in a more equal distribution of personal property among petitioners , even if the inequalities identified in broader research still hold for D.C. households.¹²⁶ Without comparable personal property data for non-petitioners in D.C., adjudicating between these alternative explanations is impossible.

In bankruptcy, debtors report personal property in 35 individual categories (see Table A4 in Appendix A for detailed definitions of each category). Because different exemptions rules apply for each type of property, examining the distribution of filers’ personal property across these categories is important. Differences in the kind and amount of property owned at the time of a bankruptcy filing directly impact how much property debtors can exempt. Figure 3 provides information on the types of personal property Black and white filers tend to claim. The figure shows the percentage of filers who report non-zero property values for a given personal property category.

Figure 3

Estimated Proportion of Filers Reporting Any

Personal Property by Type and Race

jveewheiclrley , __. …_ ◊ •

◊ • • -+-

collectibles r—- ◊

retirement ◊ •• ..

~curity deposit –i=: ◊

hobby • «>

stock •

insurance •o – •

ip g

Note: Categories of personal property are further described in Table A1 in Appendix A.

Figure 3 suggests the following two substantive conclusions. First, the most frequently reported types of personal property are clothing; household goods (which includes electronics); checking, savings and other financial accounts; vehicles;¹²⁷ jewelry; and cash. Over half of filers report ownership of each of these categories individually. Next in frequency are collectibles (e.g., antiques, stamps, and art) and retirement accounts (IRA, employer-based, Keogh, and others), which are reported by more than roughly a quarter of filers. All other types of property are reported by fewer than a quarter of filers.

Secondly, the estimates also suggests that for the most frequently reported property categories—those reported by at least a quarter of filers—Black filers tend to be slightly more likely than white filers to report a given property type. Checking and other accounts as well as cash on hand are exceptions, with white filers reporting property ownership at greater (accounts) or similar (cash) rates. For less prevalent property categories— those that are held by fewer than a quarter of filers—the opposite pattern tends to hold, with white filers being slightly more likely to report those property types.

Figure 4 provides information on claimed property values among filers. The left panel provides median property values for those reporting nonzero values of that property. The right panel provides the average share of filers’ total claimed property value that the property type represents. The sample for this panel is all filers. For example, an estimate of 30 for household goods would indicate that, on average, household goods make up 30% of all personal property value the average petitioner brings into bankruptcy with them. For both panels, only categories of property for which at least 20% of filers report owning the property type are shown.

Figure 4

Estimated Median Property Values (Left) and Estimated Share of Total

Reported Personal Property (right) by Type and Race

• •

•

Note: Categories of personal property are further described in Table A1 in Appendix A.

Estimates from Figure 4 suggest that, by value, vehicles and retirement accounts together make up the largest average value of total property reported, for both Black and white filers. For all other types of goods, median reported values tend to be low (typically under $1,000). Trends across Black and white filers indicate that more of Black filers’ property is held as vehicles (23% for whites, 33% for Blacks) and, to a lesser extent, as retirement accounts (13% for whites, 15% for Blacks) and apparel (5% for whites, 8% for Blacks).¹²⁸ With the exception of these three categories plus household goods and checking or cashing accounts, most property types make up a small share of total reported personal property values (less than 5% each).

Black filers tend to report greater median property values for vehicles ($6,815)¹²⁹ and retirement accounts ($7,542) than white filers ($5,850 and $4,889, respectively).¹³⁰ Otherwise, reported property values are similar for Black and white filers across all property types.

Estimates from Schedule B, then, suggest the following:

1. The two most valuable types of personal property

Washingtonians report owning at the time they file bankruptcy are vehicles and retirement assets. The reported median values among filers who own these types of personal property are relatively high ($6,595 in vehicle assets and $7,123 in retirement assets), especially when compared with other types of property. And a relatively large portion of filers hold these types of personal property (70% of filers report vehicle assets and 31% report retirement assets).

2. Black filers report having higher median retirement asset

values (54% more) and vehicle values (16% more) than white filers.

3. The proportion of filers who hold personal property as

household goods, financial accounts, and apparel is large (more than 80% for each category). Yet these constitute a relatively low proportion of total personal property reported for the average filer (below roughly 15% for each category individually). The median property value for each category is also relatively small (less than roughly $1,000 each category).

Overall, these trends suggest that, for D.C. filers at least, vehicles and retirement assets may play an important role in the bankruptcy process and could potentially account for some racial differences in retained assets after bankruptcy.¹³¹ Other categories of personal goods, however, are unlikely to play a major role given the small values petitioners hold for these property types. The next Section explores this in greater detail by focusing on claimed exemptions.

2. Personal Property Exemptions Claimed on Schedule C

In chapter 7 cases, debtors with personal property that exceeds any applicable exemptions may keep only their exempt property when they complete their bankruptcy case; any property beyond the exemption limits is used to repay creditors. The amount of personal property debtors may exempt varies by type of property, by filing status, and by the exemption regime claimed (i.e., federal versus D.C.).

Schedule C is where filers claim the property they intend to exempt (i.e., keep). At the median, Black filers in the District of Columbia claimed $8,359 in personal property exemptions and white filers claimed $10,150.¹³² The median white filer, in other words, sought to retain roughly $1,800 more of their property than Black filers, despite reporting similar overall personal property values on Schedule B and despite Black filers reporting 54% higher median retirement asset values and 16% higher vehicle values than white filers.

The left panel of Figure 5 shows what proportion of filers claimed exemptions for each of the allowable categories of personal property. The right panel of Figure 5 provides the ratio of the proportion of filers claiming an exemption to the proportion of filers claiming property for each property type. For example, an estimate of 0.8 for a given property type indicates that 80% of petitioners who reported nonzero property value for this type of property also claimed an exemption. For both panels, only property types where at least 20% of filers report non-zero values are shown.

Figure 5

Estimated Proportion of Filers Claiming Exemption (Left) and

Estimated Ratio Of Exemption Claims to Property Claims (Right) by

Type and Race

•

• •

cash ., cash ◊ – •

• _J

Note: Categories of personal property are further described in Table A4 in Appendix A.

The left panel of Figure 5 indicates that, unsurprisingly, filers mostly claim exemptions in those same property categories they report assets; the rank order for types of property by proportion of filers claiming an exemption and by proportion of filer reporting the property in the first place are nearly the same. For most property types, the proportion of filers claiming exemptions is slightly lower than the proportion claiming property, but for all property types in the figure at least 80% of asset holders claim an exemption. For most categories, Black filers are either slightly more likely than or as likely as white filers to claim an exemption, with the notable exception of retirement assets where all white filers claim exemptions, but roughly 5% of Black filers do not. Black filers are also less likely to claim apparel and jewelry as exempt.

Focusing on the key property categories highlighted in the previous section, the estimates in the right panel of Figure 5 suggest that roughly 60% of filers claimed vehicle exemptions.¹³³ This is true for both Black and white filers. This result diverges substantially from those reported by other researchers.¹³⁴ Nearly all Black filers (95%) who report retirement accounts also claim an exemption for it, and all white filers claiming retirement assets also claimed exemptions. Similarly, roughly 80% of both Black and white filers claim an exemption for financial accounts, for household goods, and for apparel.

The left panel of Figure 6 provides median exemption values claimed by filers with nonzero exemption claims, again focusing on the most frequently claimed property types. The right panel of Figure 6 shows the average proportion that each type of exemption represents (out of total exemptions claimed) across all filers. Total exemptions were obtained by summing all individually reported exemptions on Schedule C.

Figure 6

Estimated Median Exemptions (Left) and Estimated Share of Total

Reported Exemptions (Right) by Type and Race

• •

•

Note: Categories of personal property are further described in Table A4 in Appendix A.

The left panel of Figure 6 shows that exemptions claimed are relatively small—under $1,250—for all but two property types: vehicles and retirement accounts. Of these two, the average claimed exemptions for retirement accounts are much larger. The right panel of Figure 6 suggests that most exemptions by value pertain to vehicles, retirement accounts, and household goods, followed by exemptions for security deposits, financial accounts, and apparel. Black filers tend to hold either more of their exemptions as retirement, vehicle, household goods, and apparel, while white households disproportionately hold exemptions in security deposits and cash account—although the overall racial differences are small. For Black filers, household goods make up on average 19% of all claimed property exemptions by value, compared to 14% for white filers. For apparel, the corresponding estimates are 9% (Black) and 6% (white). For nonretirement accounts, the estimates are 6% (Black) and 11% (white).

Retirement assets are the largest exemption category (by dollar value). The proportion of all petitioners claiming an exemption in a retirement account is similar for Black and white debtors, at roughly 30% for both groups. This is surprising because in general white, non-Hispanic Americans are much more likely to have a retirement savings account (68%) than Black Americans (41%).¹³⁵ Out of petitioners who report any retirement assets, all white filers claim an exemption, while 95% of Black filers do.

Exemptions claimed for retirement accounts were larger than other exemptions in our sample for other property. As reported in Table 4, Black filers claimed roughly $7,543 in exemptions at the median for retirement accounts, compared with roughly $4,889 for white filers. Black filers, then, claim roughly $2,650 more in retirement asset exemptions than white filers.¹³⁶ The claims exactly match median reported asset amounts, meaning that at the median both Black and white petitioner exempt all retirement assets, and the observed difference in exemptions is fully explained by the difference in assets reported.

This Black-white difference in retirement account exemptions is notable because the gap in retirement assets we find is in the opposite direction relative to the gap that exists outside of bankruptcy.¹³⁷ Outside of bankruptcy, Black filers had between three and four times smaller retirement account values than white households.¹³⁸ By contrast, in our sample, Black households enter (and leave) bankruptcy with more retirement assets. Table 4. Estimated Median Retirement Account Values and Median Exemptions

Black- Black- White White

Black White Gap Ratio Assets $7,543 $4,889 −$2,654 1.54 Reported Exemptions $7,543 $4,889 -$2,654 1.54 Claimed

Vehicle assets are the other property type most commonly held and claimed as exempt. As reported in the previous section, Black filers report approximately $6,815 in vehicle assets at the median, but only seek to exempt $2,728. By contrast, white filers report lower vehicle values ($5,850) but claim nearly identical exemptions ($2,575). Put differently, white debtors claim a larger percentage of their vehicles’ value as exempt and as a result retain similar value as Black filers in absolute terms, despite holding lower valued assets.¹³⁹

Table 5. Estimated Median Vehicle Value and Median Exemption

Black- Black- White White

Black White Gap Ratio Assets Reported $6,815 $5,850 −$965 1.16 Exemptions $2,728 $2,575 −$153 1.06 Claimed

Estimates of exemptions claimed on Schedule C, then, suggest the following:

1. White filers claim approximately $1,791 more in total

exemptions than Black filers at the median ($28,359 at the mean).

2. Retirement accounts make up a large portion of retained

property value for filers (median claimed exemption of $7,124). And 31% of filers claim an exemption in a retirement account.

3. Vehicle exemptions are claimed by 58% of filers, and their

exemption claims are also sizeable ($2,744 at the median).

4. Exemptions claimed show relatively few racial differences

in retained value across the different categories of goods and assets, especially relative to the value of assets held. Racial differences are greatest at the median for retirement exemptions ($6,595 for Black filers compared to $4,889 for white filers).

5. Exemptions claimed for retirement accounts and vehicles

are higher for Black households than white households. These trends are notable in that they are opposite in direction to the disparity that exists outside of bankruptcy, where white households hold more assets on average, and opposite to the overall trend for total personal exemptions in our sample, as noted above.

6. Given the low value of exempt property, the data suggest

that the median bankruptcy filer is a struggling person or household, not an abuser of the bankruptcy system.¹⁴⁰

Conclusion

This is the first law review article (to our knowledge) to empirically examine bankruptcy exemptions by the debtor’s race, which builds on Professor Dickerson’s theoretical work on these issues.¹⁴¹ To identify a debtor’s race, we created a novel dataset based on a sample of bankruptcy petitions in the District of Columbia for 2011 and imputed petitioner race using surname and zip code.

The lack of racial identifiers in bankruptcy filings or administrative data pertaining to bankruptcies limits the scope of research on race-neutrality in bankruptcy law. Traditionally, some governing bodies did not prioritize collecting consistent and reliable racial data, for example the Social Security Administration and Centers for Medicare & Medicaid Services. Unfortunately, this lack of prioritization turns a blind eye to the role systemic racism has played in our institutional structures and individual biases and limits the research efforts to explore the concern further. Conversely, there are instances when individual-level characteristics are not collected—allegedly to mitigate any repercussion of biases in decision making.¹⁴²

The inability to identify individual-level characteristics that may impact one’s economic or social well-being inhibits researchers’ ability to adequately assess biases or unintended consequences of laws or policies on subsets of the population. Moreover, given the historical context of racial and ethnic oppression, the lack of administrative data on race and ethnicity is particularly disconcerting as it makes it difficult to assess progress regarding racial oppression and discrimination.

We report several novel results, which relate to both real and personal property. Regarding real property ownership, Black bankruptcy filers in our sample are more likely than white filers to disclose owning real property (often homes) at the time they file for bankruptcy. This result differs from other studies of homeownership rates in the District of Columbia (and nationwide).¹⁴³ As for the value of those homes, there are notable differences between the median value of homesteads owned by Black and white bankruptcy filers at the time of their bankruptcy filing. White bankruptcy filers report median homestead values ($585,000) that are $333,400 larger than the median homestead value of Black bankruptcy filers in our sample ($251,600).¹⁴⁴

The choice of applicable exemption law also mattered. Bankruptcy filers of any race who claim the D.C. bankruptcy exemptions, which allow a filer to keep all their home equity, report $34,250 greater homestead values at the median than filers that claim the federal exemptions, which are more limited.¹⁴⁵ Racial differences within our sample reveal a wide gulf between Black and white filers that is masked by a focus on aggregate statistics. When claiming the D.C. homestead exemption, Black bankruptcy filers claim homesteads that are $9,930 higher at the median, but white bankruptcy filers claim homesteads that are $110,250 higher.¹⁴⁶

Regardless of which exemptions filers selected, there were notable differences between the median claimed homestead exemptions of Black and white bankruptcy filers. Black bankruptcy filers exempted $100,250 more when asserting the D.C. exemptions ($112,000) instead of the federal exemptions ($11,750).¹⁴⁷ By contrast, white filers reported a $317,937 difference at the median between the federal exemptions ($45,463) and the D.C. exemptions ($363,400).¹⁴⁸ In short, our facially neutral bankruptcy laws facilitate white filers leaving bankruptcy with greater wealth than Black filers.

Regarding personal property, we find that Black filers report roughly similar median personal property values at the time of their bankruptcy filing. This is surprising, because earlier research led us to expect that Black bankruptcy filers would own fewer assets to begin with. Somewhat less surprising was that the median white bankruptcy filer sought to exempt roughly $1,800 more personal property relative to Black filers, even though they entered bankruptcy with similar personal property assets.

However, there were some notable differences across assets in the size of exemptions claimed. At the median, Black bankruptcy filers held and exempted more retirement assets (roughly 54% more) than white filers. The disparity for retirement assets is striking given that overall Black filers exempt less personal property than white filers, and that nationally Black households tend to hold much fewer retirement assets than white households. Black and white filers claimed similar exemptions for vehicles, which is notable given that at the median, Black bankruptcy filers report higher vehicle asset values going into bankruptcy.¹⁴⁹ As a result, facially neutral bankruptcy laws allowing filers, whether Black or white, to exempt certain personal property may facilitate white filers leaving bankruptcy with more wealth than Black filers.

Our ability to generalize from our data to the rest of the country is limited given its geographic focus on Washington, D.C. In future work, we hope to obtain a larger sample to determine whether our results hold over time and in other jurisdictions. It is vital to understand how existing social and legal structures relate to the racial wealth gap, particularly because the racial wealth gap has not narrowed substantially in the last seventy years and is reminiscent of the Jim Crow-era wealth gap.¹⁵⁰

* * *

Appendix a: Additional Tables

Tables A1 reports mean and median values of debtors’ homesteads and homestead exemptions by racial groups and type of exemption pursued (D.C. versus federal).¹⁵¹

Table A1

Reported Homestead (Schedule A), by exemption type given Race

Table A2 Reported Homestead Exemption (Schedule C), by exemption type given

Race (Filers with non-Zero Exemptions only)

1T’aa1bblleie Aa 33 l b

Differences of Means by Race 0 u

Means !Medians !Number of IMeans !Medians Numberof p-value d

I I o Obsenrations Obsenrations (two- n tailed test d of mean e egualityj_ c

Income $3,223 $2,820 303 $4,043 $3,069 49 0.115 Expenditure $3,554 $3,109 306 $4,561 $3,496 54 0.046

W Assets $170,366 $30,767 310 $225,405 $24,854 55 0.244 hi Debt $25 1,427 $11 8,389 310 $385,555 $133,392 55 0.111 Schedule A- $286,653 $251,600 156 Homestead 1$589,174 1$585,000 119 10.000 Schedule C $56,613 $9,357 307 Total $89,096 $11,450 55 0.257 1 Homestead $1091469 $211625 99 $285,451 $259,406 13 0.072 ¹⁵²/ A-1:rparel $638 $400 269 $511 $400 49 0.141 3¹⁵³2 ,— — — — ewelry $385 $200 197 $268 $200 35 0.033 5:

— — 42 Retirement $26,001 $7,543 101 Savings 1$15,808 1$4,889 117 I0.153 2 Cash $128 $40 183 $54 $40 35 0.166 0 r _ Checking or $889 $300 242 r Savings k Account $1,200 $600 48 0.264 _ r Vehicle $3,496 $2,728 189 $4,214 $2,575 29 0.479 1 d Schedule B $27,273 $11,276 312 o Total $21,301 $11,923 56 0.378 m ( Apparel $614 $400 303 $536 $350 52 0.403 i n ewelrv $365 $199 219 $288 $200 36 0.221 n Retirement $25,249 $7,543 107 o e Savin1rs $15,808 $4,889 17 0.174 n Cash $131 $34 201 $510 $40 39 0.423 Checking or $915 $297 264 r Savings e Account $1,486 $610 54 0.074 & Vehicle $8,799 $6,81 5 229 $9,928 $5,850 36 0.682 l

Table A4

Types of Personal Property Reported

Label Used in Type of Personal Property as Described in Schedule B Figures Cash on hand. cash Checking, savings or other financial accounts, account certificates of deposit, or shares in banks, savings and loan, thrift, building and loan, and homestead associations, or credit unions, brokerage houses, or cooperatives. Security deposits with public utilities, telephone security companies, landlords, and others. deposit Household goods and furnishings, including audio, household video, and computer equipment. goods Books, pictures and other art objects, antiques, stamp, collectibles coin, record, tape, compact disc, and other collections or collectibles. Wearing apparel. apparel Furs and jewelry. jewelry Firearms and sports, photographic, and other hobby hobby equipment. Interests in insurance policies. Name insurance insurance company of each policy and itemize surrender or refund value of each. Annuities. annuity Interests in an education IRA as defined in 26 U.S.C. § education IRA 530(b)(1) or under a qualified State tuition plan as defined in 26 U.S.C. § 529(b)(1). Give particulars. (File separately the record(s) of any such interest(s). 11 U.S.C. § 521(c).) Interests in IRA, ERISA, Keogh, or other pension or retirement profit-sharing plans. Give particulars. Stock and interests in incorporated and unincorporated stock businesses. Interests in partnerships or joint ventures. Itemize. partnership Government and corporate bonds and other negotiable bonds and nonnegotiable instruments. Accounts receivable. receivable Alimony, maintenance, support, and property alimony settlements to which the debtor is or may be entitled. Give particulars. Other liquidated debts owed to debtor including tax liquidated refunds. Give particulars. debt Equitable or future interests, life estates, and rights or future interest powers exercisable for the benefit of the debtor other than those listed in Schedule A–Real Property. Contingent and noncontingent interests in estate of a inheritance decedent, death benefit plan, life insurance policy, or trust. Other contingent and unliquidated claims of every claims nature, including tax refunds, counterclaims of the debtor, and rights to setoff claims. Give estimated value of each. Patents, copyrights, and other intellectual property. IP Give particulars. Licenses, franchises, and other general intangibles. Give license particulars. Customer lists or other compilations containing customer list personally identifiable information (as defined in 11 U.S.C. § 101(41A)) provided to the debtor by individuals in connection with obtaining a product or service from the debtor primarily for personal, family, or household purposes. Automobiles, trucks, trailers, and other vehicles and vehicle accessories. Boats, motors, and accessories. boat Aircraft and accessories. plane Office equipment, furnishings, and supplies. office

equipment Machinery, fixtures, equipment, and supplies used in machinery business. Inventory. inventory Animals. animals Crops–growing or harvested. Give particulars. crops Farming equipment and implements. farm

equipment Farm supplies, chemicals, and feed. farm supplies Other personal property of any kind not already listed. other Itemize.

Appendix B: The Dataset

This Article uses an original dataset compiled from a random sample of bankruptcy petitions filed in 2011 in the United States Bankruptcy Court for the District of Columbia. We use a sample of 501 bankruptcy petitions, out of a total of 959 petitions filed that year.¹⁵⁴ Jialan Wang, Assistant Professor of Finance at Gies College of Business at the University of Illinois Urbana-Champaign, originally obtained the petitions from the Public Access to Court Electronic Records (PACER) website as part of an unrelated project.¹⁵⁵ Professor Wang generously provided us access to the bankruptcy petitions themselves, as well as to summary of schedules and all debtors’ schedules (including any amended schedules).¹⁵⁶ Other documents were also occasionally available.

We opted to analyze 2011 bankruptcy filings because they were the most recent filings made available to us that also occurred within a period that we deem roughly representative of the overall D.C. bankruptcy trends. In particular, the 2011 filing occurred substantially after the last major revision to the Bankruptcy Code (the 2005 amendments, which dramatically altered filing patterns in 2004–2006)¹⁵⁷ and a few years after the Great Recession (which was also associated with atypical filing patterns).¹⁵⁸

The District was chosen as the main geographic focus due to demographic concerns. D.C., affectionately known as “Chocolate City” since the release of the song by the band Parliament in 1975, has a population that is currently 41% Black. Since 1973, the city has also exclusively had Black mayors. The Black political leadership and substantial Black population mean that bankruptcy laws in D.C. could come close to a “best case scenario” in terms of bankruptcy designs that do not exacerbate racial inequities. As such, we considered D.C. a conservative test case for our purposes. Two of the authors also reside in D.C., which helped contextualize some of the nuances of the data.

When filing for bankruptcy, debtors need to provide a substantial amount of personal (e.g., name, address, Social Security number, etc.) and financial (e.g., assets and liabilities, monthly income and expenses) information.¹⁵⁹ For this project we focused in particular on the reported value of the debtor’s real and personal property at the time the bankruptcy case was filed and on the property debtors claimed as exempt.¹⁶⁰

A. Data Entry Work

Generally, the entire set of documents available for each debtor was between thirty-eight and fifty-two pages. Because the petitions sampled came from the same court in the same year, these PACER documents generally followed a similar structure. These documents, however, were typically only available as PDF scans of original handwritten petitions or PDF copies of electronically submitted documents. As a result, a significant part of our work consisted in converting the PACER petitions to a machine-readable spreadsheet format.

Because the order in which individual sections of documents were included varied across individual PDFs, and because petitions usually included multiple PDF files (including amended documents), contained handwritten information, and did not always include the same set of elements (e.g., a debtor’s attorney sometimes included a disclosure of attorney compensation as the first page in the file), the data-coding process could not be automated.

Under the supervision of Matthew Bruckner, two research assistants were hired from Howard University School of Law’s second year class to convert each bundle of PDF documents to a spreadsheet-format dataset.¹⁶¹ Coders were asked to retrieve information for up to eighty-nine variables per debtor. Professor Bruckner established coding guidelines and reviewed them with students. Students then individually coded five petitions. Professor Bruckner reviewed their work and guidelines were amended as appropriate. For the remaining petitions, students encountering issues not covered in the guidelines would request feedback on a one-off basis, and updated guidelines would be shared with the entire team as needed.

From each petition, coders collected information about the debtor (e.g., name and address) and about their bankruptcy case (e.g., whether the debtor had filed a previous bankruptcy, whether the debtor was represented by an attorney or assisted by a petition preparer, which chapter of the bankruptcy code the case was filed under, and whether the case involved primarily business or consumer debts). Coders also collected information from the statistical summary of certain liabilities, the summary of schedules, and Schedules A, B, C, F, I, and J. ¹⁶²

Coders used Schedules A and B to collect debtors’ financial information, and in particular, information on assets held at the time of filing. For example, coders collected information on checking account ownership and, if the debtor owned a checking account, coders recorded the amount of money in that account. Coders also collected information on a variety of other assets, such as household goods, furniture, clothing, appliances, books, animals, crops, and jewelry. If debtors had these assets, coders entered the value the debtors listed for such items on their schedules.

Coders used Schedule C to collect information on debtors’ claimed exemptions. Coders used the description of the item provided by debtors to code the exemption into one of the thirty-five categories of assets listed in Schedule B (see Table A4 for list), and then listed the value of the claimed exemption. For example, a debtor might enter bankruptcy with a 1992 Toyota Camry with a fair market value of $2,375 listed in their Schedule B. Under 11 U.S.C. §522(d)(2), a debtor filing for bankruptcy in 2011 was allowed to keep up to $3,225 in equity in one motor vehicle.¹⁶³ If debtors claimed that exemption on their Schedule, the coders included it in our dataset as an “automobiles, trucks, trailers, and other vehicles and accessories” exemption. Coders also noted whether the debtor claimed the District of Columbia or federal exemptions.

Although not used in this Article, coders also collected some information on the nature of the debtor’s unsecured debts and on reported expenditures.

To promote accurate data entry, each coder independently coded all 501 petitions in our sample. For handwritten petitions especially, values inputted by petitioners could sometimes be difficult to ascertain. Once each coder had individually collected all the relevant data, Raphaël Charron-Chénier compared duplicate entries for consistency and returned nonmatching entries to coders for recoding against the original petitions. An initial comparison of debtor’s names (which was necessary to match entries across duplicate files) found approximately 110 nonmatches across 501 debtors. These were due primarily to minor typographical errors or inconsistent use of middle names and other designations (e.g., Jr.). Given the importance of debtors’ names for our project,¹⁶⁴ both coders checked every nonmatching petition against the original record and made corrections where necessary. We repeated the process until we had 100% agreement by both coders on all debtors’ names.¹⁶⁵

Once the duplicate datasets could be matched by debtor name, the remaining variables were compared. That comparison revealed mismatches on 3,778 individual data cells out of a total of 23,700, a roughly 16% error rate. Most mismatches appear to be the result of typographical errors or arose because of unclear information on the petitions themselves.¹⁶⁶ Mismatches were distributed roughly evenly across debtors, with nearly every debtor (480 of 501) having a mismatch across at least one variable coded. Half of the entries had mismatches across four or fewer variables, and 90% had mismatches across fewer than ten. As before, each mismatched data cell was flagged, and each coder was instructed to compare entries against the original record and make a correction if appropriate.¹⁶⁷

Some variables we had initially planned to use in analyses proved to have particularly high coder error rates. For instance, coders were asked to identify the amount of medical debt listed on debtors’ Schedule F. Some debtors describe medical debt clearly, listing specific liabilities as “medical debt.” Other entries, however, might list a veterinary hospital or a medical device as the liability source. Coders’ judgements sometimes varied as to whether the liability should be included as medical debt. Out of concern for the accuracy of our dataset, we ultimately opted not to include variables on student, medical, and credit card debt that were derived from information provided in Schedule F.¹⁶⁸

Additionally, the authors re-reviewed all the underlying documents related to retirement savings after our initial analysis suggested that there might be data anomalies. For example, in an earlier draft, we reported that the median Black debtor was not exempting all their retirement savings, despite having the legal right to do so.¹⁶⁹ This process uncovered a few errors, which we corrected for this draft.

B. Constructing a Race Variable and our Proxying for Race

Bankruptcy petitions do not require debtors to identify themselves by race or ethnicity. To determine a likely racial identifier for debtors, we use the Bayesian Improved Surname Geocoding (“BISG”) approach.¹⁷⁰ The BISG uses surname and geographic location to estimate a racial proxy. Census data is used to obtain an aggregate measure of racial and ethnic identity for a list of roughly half of all U.S. surnames (151,671 surnames totaling about 294 million individuals).¹⁷¹ This information is combined with Census data on the racial and ethnic composition of the population across five-digit zip codes.¹⁷² The information is combined using Bayes’ theorem to obtain a probability that a given debtor identifies as being of a given racial or ethnic group.

The BISG was initially developed in 2009, and in 2014 the Consumer Financial Protection Bureau (the “CFPB”) published a methodological brief on “Using publicly available information to proxy for unidentified race and ethnicity.”¹⁷³ The BISG method improves on previous approaches to imputing race and ethnicity by using both surname and zip code methods simultaneously. Elliott et al., (2009) found that BISG affords a 41% efficiency gain compared to surname-only methods, and an over 100% efficient gain relative to geocoding-only techniques.¹⁷⁴

For each debtor, the BISG provides a probability estimate of racial identification. For our analyses, we created a race proxy using a 50% probability threshold: debtors were assigned a proxy race or ethnicity only if their probability of belonging to a given racial or ethnic group was estimated to be at least 50%. In our sample, after excluding business-related filings, 13 petitions did not meet this 50% threshold and were therefore not retained for analyses. We were unable to use a higher threshold, as some other researchers have done,¹⁷⁵ given the limited number of filings in our sample. Using a 90% threshold, for example, would have required excluding nearly half of the petitions from analysis.

Using this method, we estimate that roughly 74% of debtors in our sample are Black, 14% are white and 12% fall into another category.¹⁷⁶ Proxy race estimates for the full sample of non-business filers can be found in Table B1, below. Our analyses focus on Black and white filers only. Estimates from the 2010 U.S. Census Bureau indicate that 37.5% of D.C. residents are (non-Hispanic) white and 45% are (non-Hispanic) Black.¹⁷⁷ White filers are therefore underrepresented and Black filers overrepresented relative to the D.C. population.¹⁷⁸ This is consistent with previous research that finds Black debtors tend to be overrepresented among bankruptcy filers relative to their share of the population. The disparity in our sample (one and a half times) is smaller than that reported by other researchers, however, who note that Black filers tend to be overrepresented in bankruptcy by a factor of two to three times.¹⁷⁹

Table B1

Proxy Race Estimates using BISG

R-a-c-e -or- E-t-hn-i-ci-ty- -5-0%- -T-hr-e-sh-o-ld- – – – – – – – 7-0%- -T-h-re-sh-o-ld- – – – – – –

N Proportion | p < 0.5 N Proportion | p < 0.7 Black 326 0.74 286 0.74 White 60 0.14 45 0.12 Other 54 0.12 52 0.14 Below threshold 13 – 70 – Total 440 1 384 1

C. Data Limitations

Given the richness of our data relative to other available bankruptcy datasets—in particular, our ability to examine assets and exemptions both in the aggregate and for particular categories of goods and services—our primary aim in this project is to provide a descriptive overview of petitions for D.C. residents who filed for bankruptcy in 2011.¹⁸⁰ Our conclusions, however, are limited to this particular sample. We avoid overarching claims about D.C. bankruptcies in other years or bankruptcies in other jurisdictions.

Bankruptcy petitions suffer from occasional missing data issues. Filers occasionally leave off some information requested by bankruptcy forms. Generally, we treat blank entries as a missing value, and therefore exclude these data points from the specific estimates that would make use of them (i.e., we use pairwise rather than listwise deletion). To help readers understand where missing cases were excluded, we report sample sizes used for different estimates; see, for example, Table A3 in Appendix A. For some specific estimates from Schedules B and C, blank entries on petitions are generally used to indicate the absence of an asset or the failure to claim an exemption. For these schedules, we coded blank entries as valid zeros. If some entries were left blank as an oversight rather than to indicate lack of assets or exemptions, this could potentially bias our estimates downward.

Bankruptcy filers are likely to commit some errors and omissions (involuntary or otherwise) when preparing their petitions. As bankruptcy cases move forward, these errors or omissions may be corrected. These corrections, however, may not be reflected in our data. Such entry errors by the original petitioner may account for slight discrepancies we observed across some estimates. Given our limited sample size, we opted not to drop petitioners with inconsistent answer patterns. Instead, we indicate via footnote whenever an apparent data discrepancy arises because of likely filer error and, where possible, exclude that data point from relevant analyses. Given the legal ramifications of dissimulating information on these petitions, we believe most errors are unintentional. Given the incentives filers have to report correctly and the large number of filers with bankruptcy counsel, we also believe errors are likely limited in numbers.