Serial Bankruptcy A 20-Year Study of Utah Filers

Journal Issue

Feb 2006

Column Name

Bankruptcy by the Numbers

Citation

ABI Journal, Vol. XXV, No. 1, p. 24, February 2006

Topic Tags

Consumer Bankruptcy

Fraud

Journal HTML Content

A 1996 survey of bankruptcy professionals by the American Bankruptcy Institute2
reported that two-thirds of the creditor representatives and nearly half of
the judges viewed serial filings as a significant problem. In October 2001,
the U.S. Trustee Program launched the Civil Enforcement Initiative to improve
the effectiveness of bankruptcy administration. Among other concerns, serial
filers and substantial abuse were identified as issues that present significant
challenges to the bankruptcy system.3 While serial filers have been identified
as a problem, there has been little research on these debtors.

Sullivan, Warren and Westbrook4 found that about 8 percent of their sample
of debtors had filed within the previous six years. Norberg5 reported that 28
of 71 debtors (39 percent) in the Southern District of Mississippi listed a
previous case on their petition. Debtors who listed only one prior case in the
previous six years were significantly more likely than first-time debtors to
achieve a discharge. Norberg concluded that filing a single previous case was
not an indicator of abuse, but that multiple filings were potentially abusive.
 Strand, Hira and Carter6 studied the demographic and financial attributes
of repeat debtors. They compared bankruptcy filers in one Canadian district
to one U.S. district and found that 14 percent of debtors in each location reported
a prior case. Compared to first-time debtors, repeaters were older and had lower
incomes, lower assets and lower debt levels.
 The Utah Study
 Our present study of multiple filers was designed to:

• determine the extent of serial filings in Utah from 1984 to 2004, 
• estimate the extent of debtor abuse by serial filers, 
• establish a financial and demographic profile of debtors who appear
to be abusing the system, and 
• develop a statistical model to compare demographic and financial variables
of potential abusers to nonabusers.
 This study is unique because it utilizes a 20-year time frame and a much larger
sample than prior studies that only considered cases filed within the previous
six years. Using the PACER/Electronic Case Filing System, previous filings were
tracked by Social Security number. Using Social Security numbers reduces but
does not completely eliminate the problem of incomplete information on prior
filings because debtors could have filed previously in other districts.
 The incidence of repeat filers in Utah was measured to estimate the extent
to which repeaters may be abusing the bankruptcy system. Utah has the highest
filing rate as measured by number of households (36.5) per case7 as well as
a high percentage (40 percent) of chapter 13 cases at the time the reference
cases were filed in 1997.8 This year was selected in order for sufficient time
to elapse to determine the outcome of chapter 13 cases. Nationwide, about 30
percent of filings are chapter 13 cases. Districts with low chapter 13 rates
are unlikely to have many serial filers due to the time limit between chapter
7 discharges. About one-third of confirmed chapter 13 plans are completed and
the remaining unsecured debts discharged, but this rate varies considerably
among districts.9 Only 26.5 percent of 1997 chapter 13 filers in Utah completed
their plans.10

Because the data were limited to the information in the court records and
individual circumstances were unknown, abuse was characterized only by the number
of times a debtor filed. The 20-year period (1984-2004) was used because it
is the longest time frame available on PACER. Only debtors who filed in the
same status (always joint with same spouse or always single) were included in
the study to eliminate possible marital status changes that might justify a
repeat filing.
 Methodology
 A conceptual framework was developed to sort repeat filers into categories
based on the number of filings and the time frame during which these cases occurred.
Debtors with three filings within two years and debtors with four or more filings
during the 20 years were classified as serial filers and potential abusers.
(See Figure 1).
<center><img src="/AM/images/journal/bbtnchart12-06.gif"></center>
 After a random start, cases were systematically selected from the 12,055 debtors
who filed for bankruptcy in Utah in 1997. Financial, employment, family and
demographic variables were collected from the court records using PACER.
 Based on the conceptual framework, all cases with more than one filing were
categorized as either a serial filer (potential abuser) or repeat filer (not
abuser) based on the number and pattern of filings. Thus, the dependent variable
consisted of these two categories. Independent variables included demographic
and financial characteristics from the bankruptcy files: 1997 chapter, filing
status, marital status, household size, secured debt, unsecured debt, monthly
income and monthly expenses.
 Descriptive statistics were used to compare the two groups of debtors. Analysis
continued by establishing bivariate relationships between the independent variables
and the dependent variable (serial or repeat filer). The t tests were
used to compare the two groups on continuous variables including income and
debt levels. Chi-square analyses evaluated whether serial (abuser)/repeat (nonabuser)
status is independent of each categorical variable. Because the dependent variable
(serial or repeat) was dichotomous, logistic regression was used. The logistic
regression model was built using a “backward elimination” approach,
starting with the full model containing all independent variables. At successive
steps, new models were computed by eliminating the independent variable that
was the least significant in its association with the dependent variable: serial
vs. repeat filers. The final model contained all independent variables that
remained significant in the presence of all other variables in the model.

Results
 Of the original sample of 2,567 cases, 373 cases were eliminated from the
analysis due to a change in filing status, resulting in 2,194 cases in the final
sample. A total of 235 cases (10.7 percent) were identified as serial filers
and possible abusers. Three-quarters (N=171; 73 percent) of these potential
abusers filed chapter 13 in 1997, and 64 (28 percent) filed chapter 7. The majority
(76 percent) of the potential abusers had their 1997 case dismissed. Only five
(2.9 percent) of the 171 serial debtors who filed chapter 13 in 1997 completed
their payment plan. Of the 64 serial debtors who filed chapter 7, 51 received
a discharge (79.7 percent).
 Using the backward-elimination approach, the full logistic regression model
began with all independent variables: chapter filed in 1997, filing status,
marital status, household size, secured debt, unsecured debt, monthly income
and monthly expenses. The logistic-regression model found 1997 chapter, filing
status, unsecured debt and monthly income to be the most significant variables
in identifying serial filers. Serial filers were nearly five times as likely
to have filed for chapter 13 in 1997 than chapter 7. Males and females filing
alone were nearly 50 percent less likely than joint filers to be serial filers.
Debtors who owed unsecured debts exceeding the median amount were less likely
to be serial filers. Debtors who had incomes above the median for this sample
were almost twice as likely to be categorized as serial filers. The logistic-regression
model correctly predicted 89.3 percent of serial filers who were potentially
abusing the system.
 Conclusions and Discussion
 The principal finding of this study is that 10.7 percent (235 of 2,194) of
the sample filed multiple times and thus may have been abusing the bankruptcy
system. The results are consistent with the few previous studies of repeat filers.
 Another key finding is that out of the 171 serial potential abusers who filed
chapter 13 in the reference year of 1997, only 5 (2.9 percent) of them completed
their repayment plan. Serial filers who previously discharged their debts in
chapter 7 may be filing chapter 13 repeatedly, with no intention of completing
a repayment plan, until they can once again file a chapter 7 and discharge their
unsecured debts. Also, 20.3 percent (13 of 51) of the serial debtors who filed
chapter 7 had their cases dismissed. At the time this study was conducted, debtors
who had previously received a chapter 7 discharge could not obtain another chapter
7 discharge for six years, so their only option was to file a chapter 13.
 According to the logistic regression, the only significant demographic variable
in estimating abuse by repeat filing was the filing status of the debtor. Males
and females filing alone were nearly 50 percent less likely to be abusers compared
to joint filers. While single parents often experience financial stress, they
were not over-represented in this random sample of debtors. Although changes
in marital status (i.e., divorce, remarriage, widowhood) are significant economic
events, tracking and analyzing these changes was beyond the scope of this study.
 Serial filers had higher secured debt but lower unsecured debt than repeat
filers (see Figure 2). These serial filers probably discharged debts in previous
bankruptcies, resulting in lower unsecured debt levels on their 1997 petitions.
In addition, serial filers reported higher monthly income than the rest of the
sample. Because these potential abusers were more likely to be joint petitioners,
the higher incomes may be due to two earners. The logistic-regression results
revealed that unsecured debt and monthly income were the only two financial
variables significant in identifying serial filers. Filers with unsecured debt
levels above the median were nearly 70 percent less likely to be categorized
as potential abusers; debtors with incomes above the median were almost twice
as likely to fit the definition of abuse.

<center><img src="/AM/images/journal/bbtnchart22-06.gif"></center>
 Limitations
 Because data were limited to the court records, it was difficult to determine
whether serial filers were abusing the system or had legitimate reasons for
multiple filings. It is not always clear when a bankruptcy petition is filed
in bad faith. However, it is hard to imagine a legitimate reason for one couple
to file 11 times. One male filed eight cases; one female debtor filed seven
times. When the files of these three cases were examined in detail, it was difficult
to determine whether they took on more credit obligations between subsequent
filings, or if higher debt levels were due primarily to high interest rates,
penalties, fees and collection costs. If lenders extended new credit after more
than two bankruptcy filings, these creditors are contributing to the bankruptcy
epidemic.
 Because so few petitioners in Utah have incomes exceeding the state median,
the new means test is unlikely to affect the incidence of repeat filers. The
serial filers in this study were already filing chapter 13 repeatedly and failing
to complete their proposed plans. If more debtors are forced to file 13 rather
than 7, then the discharge rate is likely to decrease. Court documents do not
provide sufficient information to make judgments about the intentions or individual
circumstances of debtors.
 Secondly, this study categorized potential abuse based solely on the number
of times a debtor filed in this district over a 20-year period. This approach
was used because of the limited information in the court records. Ideally, one
would examine each case in order to obtain more specific information on their
financial situations at each filing to determine if they appeared to be abusing
the system. However, examining all the court records may not provide a complete
picture of the debtor’s circumstances or intentions. Interviews with debtors
and attorneys might shed further light on the circumstances.
 The data are also limited to only one federal bankruptcy court district (Utah).
Debtors who moved into or out of the state during the 20-year period could have
filed in other court districts and therefore may be incorrectly labeled as nonabusers.
 Recommendations
 Each bankruptcy court should implement policies to identify, evaluate and
monitor repeat filers. Using Social Security numbers, court officials can track
previous filings and impose greater scrutiny upon those debtors who file multiple
times. Some of these debtors may have problems such as alcohol, gambling or
other addictions that can be dealt with more efficiently outside of bankruptcy
court.

While the Bankruptcy Abuse Prevention and Consumer Protection Act of 2005
(BAPCPA) extends the period between chapter 7 discharges from six to eight years
and limits are placed on repeat chapter 13 discharges, serial filers are likely
to continue to file multiple chapter 13 cases during the intervening years.
The frequency of repeat filings is likely related to chapter 13 filing rates
in individual districts, which vary substantially. Replication of this study
with a national sample including other bankruptcy court districts, especially
those with high chapter 13 filing rates, is needed.
 Although there are undoubtedly some instances of abuse, results of this study
suggest that abuse of consumer bankruptcy by repeat filers is not extensive
and can be managed by identifying serial filers. Among the many changes in practice
resulting from the new law, attorneys should require clients to fully disclose
prior filings. Trustees should run a check on the social security numbers of
debtors reporting prior filings to see their full history to cut down on abusive
serial filers.
 Author’s Note: The researchers completed Utah State University’s
Institutional Review Board human subjects training and confidentiality measures
were taken to assure that Social Security numbers and all other personal information
remained private. The researchers thank U.S. Bankruptcy Court Chief Judge Glen
E. Clark (D. Utah) for granting access to the PACER system and waiving
fees. This study is part of a larger project examining serial filers funded
by the ABI Endowment Fund.
<blockquote>
<blockquote>  </blockquote>
</blockquote>

<hr>
<h3>Footnotes</h3>

1 Also written by Bonny Llewellyn, M.S., a chapter 13 internal auditor in Salt
Lake City. 
2 American Bankruptcy Institute (December 1996). Report on the State of
the American Bankruptcy System. Available at: www.abiworld.org/Content/ContentGroups/News_Room1/Bankruptcy_Research_C…;
 3 Darling, Antonia G. and Redmiles, Mark A., “The Civil Enforcement
Initiative: a Review of the First 10 Months and a Look at the next Stage,”
22 Am. Bankr. Inst. J. 5 (2002). Available at www.usdoj.gov/ust/eo/public_affairs/articles/docs/abi_092002.htm.

4 Sullivan, Teresa A., Warren, Elizabeth and Westbrook, Jay L., As We Forgive
Our Debtors: Bankruptcy and Consumer Credit in America. New York: Oxford
University Press (1989).

5 Norberg, Scott F., “Consumer Bankruptcy’s New Clothes: An Empirical
Study of Discharge and Debt Collection in Chapter 13,” 7 Am. Bankr. Inst.
L. Rev. 415 (1999). 
6 Strand, Judith K., Hira, Tahira K., and Carter, Richard B., “Repeat
Consumer Bankruptcy: A Comparative Analysis with One Time Petitioners in the
United States and Canada,” Proceedings of the Association for Financial
Counseling and Planning Education, 142 (1994). 
7 American Bankruptcy Institute, “Households per Filing, Rank (2004).”
Available at www.abiworld.org/statcharts/HouseRank.htm. 
8 1997 was the reference year because data had already been collected from
the case files for a random sample of cases filed that year. 
9 Bermant, Gordon, and Flynn, Edward, “Measuring Performance in Chapter
13: Comparisons Across States,” 19 Am. Bankr. Inst. J. 6. (July/August
2000). Available at www.usdoj.gov/ust/eo/public_affairs/ articles/docs/abi082000ch13.htm.

10 Evans, David A., “Predictors of 1997 Chapter 13 Bankruptcy Completion
and Dismissal Rates in Utah,” unpublished Master’s thesis, Utah
State University, Logan, Utah (2004).

Journal Authors

Jean M. Lown

Journal Date

Wednesday, February 1, 2006 - 12:00

Bankruptcy Rule

2006

none