INCOMES, DEBTS, AND REPAYMENT CAPACITIES OF
RECENTLY DISCHARGED CHAPTER 7 DEBTORS

Gordon Bermant and Ed Flynn
Executive Office for United States Trustees
January, 1999

INTRODUCTION:

      In December, 1998, the Executive Office for United States Trustees (Executive Office) completed the first stage of an ongoing study of chapter 7 consumer bankruptcies. The Executive Office is uniquely situated to do this work, in part because virtually all of the information associated with chapter 7 cases flows through the United States Trustees' offices throughout the country.

     Having reliable information is particularly important now, for two reasons. First, bankruptcy filings have increased substantially in recent years. Total consumer bankruptcy filings (filings under chapters 7 and 13) have exceeded one million per year each year since 1996. In 1998, nearly one million consumer chapter 7 cases and 400,000 chapter 13 cases will be filed.

     Second, the 105^th Congress considered several revisions to the Bankruptcy Code that would test the ability of consumer debtors to repay their debts during chapter 13 plans designed to last for five years. These tests are often referred to as "means testing". The proposed legislation would require debtors who met the tests' criteria to enter chapter 13 or risk having their cases dismissed. At the end of the legislative session, a House-Senate Conference Committee reported a bill that included tests similar to those presented in earlier House and Senate bills. Among other things, the bill required a calculation of allowed expenses based on IRS Guidelines for housing, food, transportation, and other necessary expenses. The session ended before the Conference bill came to a vote in the Senate. This bill or a similar one may well be considered by the 106^th Congress.

     Any amendment to the Code that includes a test of debtors' ability to repay will significantly affect the work of United States Trustee offices throughout the country. In the first place, U.S. Trustee staff oversee the work of the panel trustees who are charged to make the initial tests of debtors' abilities to repay. This will involve checking the accuracy and thoroughness of the panel trustees' calculations of debtors' ability to repay. In addition, the U.S. Trustees will continue to bring motions against debtors who apparently have repayment capacity, regardless of motions made by the panel trustees.

OTHER PUBLISHED STUDIES: Three other consumer bankruptcy studies have been disseminated since October 1997. The Credit Research Center released a study based on 2,441 Chapter 7 cases filed during 1996 in 13 cities.⁽¹⁾ Ernst & Young published a study including 2,220 cases drawn from all 90 judicial districts and filed during 1997.⁽²⁾ These two studies, which were funded by the credit industry, concluded that one year's worth of chapter 7 debtors could repay between $4 billion and $5.1 billion to unsecured creditors under five-year chapter 13 plans. Professors Marianne Culhane and Michaela White of Creighton University Law School have completed a study, funded by a grant from the American Bankruptcy Institute, that included 1,043 cases filed during 1995 in 7 districts.⁽³⁾ Culhane and White concluded that chapter 7 debtors could repay no more than $800 million to unsecured creditors during their five-year chapter 13 plans.

POPULATION OF THE PRESENT STUDY: The study used information gathered from the petitions and schedules filed in approximately 2,000 chapter 7 cases closed by the U.S. Trustees during the first half of 1998. Almost all of these cases had been filed during late 1997 or very early 1998. The cases were gathered from each federal judicial district in proportion to the total number of chapter 7 cases filed in each district during 1997. Appendix 1 describes the distribution of the study sample in detail. The final number of cases included was 1,955. All of the cases had been designated by the panel trustees as containing no assets for distribution to creditors.⁽⁴⁾

     During the year ended March 31,1998, there were 975,370 consumer chapter 7 cases filed nationwide.⁽⁵⁾ All but about 10,000 of these cases will be closed as no-asset cases. Thus, our sample represents about 1/500th of annual national no-asset chapter 7 filings. Characteristics of the sample can be extrapolated to estimate equivalent characteristics of the national picture of chapter 7 bankruptcies.

FINDINGS:

DEBTOR CHARACTERISTICS: One of the most striking features of the population of chapter 7 debtors is the great variability displayed across the important measures, including the three major categories of debt, gross and net income, and reported expenses. Small numbers of debtors have high incomes, or debts, or expenses. The skew in these distributions has two immediate implications for how the information should be described and inferences drawn from it. First, whenever feasible the data should be summarized by reference to their medians (mid-points) rather than by their means (arithmetic averages); and second, debtors at the high end of the distributions are strikingly unrepresentative of the great numbers of debtors.

     This feature of the population of debtors is illustrated in the following table, which displays the means, medians, and maximum values for incomes, expenses, and debts. It also shows the percentage of the sample reporting zero values in each category. The impact of extreme values is clearly illustrated by the priority debt category. Fewer than 20% of the sample had any priority debt at all, but a few debtors had very large priority debts (primarily tax debt and student loans), resulting in a mean for the entire sample of $1,525.

HOUSEHOLD SIZE: Under means-testing, the debtor's family size must be known in order to compare the debtor's gross income against the appropriate national median income. The average household size of the sample was 2.36, a little below the national average household size of 2.62. Over 60% of the study population were in households of either one or two persons.

DEBTOR CATEGORIES: Based on reported gross monthly income and the national median income standards specified by the House and Senate bills ⁽⁶⁾, we divided the sample population into the following four groups.

1. "Under Median" debtors: 1,345 (68.8%) of the debtors reported gross monthly income below both the House and Senate standards. These debtors would remain eligible for chapter 7 under the means testing formulas.

2. "Senate Gap" debtors. 247 (12.6%) of the debtors had incomes above the Senate thresholds but below the House thresholds. All but one of these debtors were in one-person households with gross monthly incomes between $1,491 and $2,325. Application of the IRS expense allowances to debtors in this category would result in nearly all of them having no available income to fund a chapter 13 plan.

3. "House Gap" debtors: 16 (.8%) of the debtors had income above the House thresholds but below the Senate thresholds. Ten of these debtors had households of five or more persons. Most debtors who fall in this category would not have surplus income after applying the IRS standards, and would therefore not be required to file in chapter 13.

4. "Over Median" debtors: 347 (17.7%) of the debtors reported gross monthly income above both the House and Senate thresholds. Under any means testing scenario proposed in last year's legislation these would be the debtors with all or most of the total repayment capacity. Therefore, these debtors were the focus of most of our analysis.

       The four categories of debtors have substantially different financial profiles. In particular, the petitions of the Over Median debtors showed about double the average debt, income, and expenses of the typical debtor.⁽⁷⁾

DEBT PROFILES: We analyzed the unsecured debts (Schedule F) of the 347 Over Median debtors, and placed them in the following eight general categories.⁽⁸⁾

     The unsecured debt statistics are severely skewed by a few debtors with extremely high unsecured debts. More than one-half of the unsecured debt of the 347 Over Median debtors was held by the 12 debtors who each owed more than $250,000. Most of this debt was the result of the operation of a business, or legal and deficiency judgments. In contrast, nearly two-thirds of the debt of the other 335 Over Median debtors with under $250,000 in unsecured debt was owed to credit card companies.

NATIONAL REPAYMENT ESTIMATES: After reducing the Over Median debtors' incomes to account for business expenses, tax liability, support and alimony payments, and priority debt payments, we were left with a pool of income from which some could go to unsecured creditors. By this measure, 300 of the 347 Over Median debtors in our study had available income.

      If all of this pool of income were used for repayment, debtors' incomes and expenses did not change, and all debtors were able to complete a five-year repayment plan, unsecured creditors would receive $3.76 billion over five years from each year's cohort of debtors.⁽⁹⁾ This figure is nearly as high as the credit industry estimates of repayment capacity. But in order to realize this amount, all debtorsmoved into chapter 13 by virtue of means testing would have to live at the national median income level for their family size. Presumably, this would entail significant changes in the lives of many of these debtors.

    If, however, instead of taking 100% of the available pool for the unsecured creditors, the system took either 75%, 50%, or 25% of it, the unsecured creditors could potentially gain $3.22 billion, $2.49 billion, or $1.40 billion, respectively.

     Some chapter 7 debtors would be able to pay all of their unsecured debt under any of the 4 plans described in the previous paragraph. The number of such debtors is small-certainly under five percent of all chapter 7 debtors. For example, in our study population 55 debtors (2.8%) could repay their unsecured creditors in full if one-half their surplus income were devoted to a repayment plan, and 16 debtors (0.8%) could repay their unsecured creditors in full if one-quarter of their surplus were devoted to a repayment plan.

FACTORS THAT WOULD AFFECT REPAYMENT ESTIMATES: Thus, compared to several credit industry estimates that means testing could result in an additional $4 billion to $5 billion per year being repaid to unsecured creditors, our initial estimates based on repayment of all or a portion of excess income range from $1.4 billion to $3.76 billion. Any of these figures, if accurate, would represent an enormous change over present experience since unsecured creditors now receive less than $1 billion per year in chapter 7 asset cases and chapter 13 cases.

    However, we believe that other variables would act to reduce repayments under means testing even further. Over a five-year period many of these debtors will experience some type of change, such as job loss or other reduction of income, divorce, remarriage, and so on. These life changes will affect either their income or expenses and thus their repayment ability. The parallels reports that only about one-third of current voluntary chapter 13 cases result in completion of a repayment plan; the others are dismissed or converted.

    IMPACT OF CREDIT COUNSELING: Section 321 of S. 1301 required consumer debtors to receive credit counseling within 90 days before filing for bankruptcy. This provision may have a substantial impact on who files for bankruptcy and under what chapter they file. Over time, it may substantially reduce the number of chapter 13 cases filed.

     UN-REPAYABLE DEBTS: Our study population contained a sizable number of debtors who, under any circumstances, had no apparent ability to repay their debts. The study included 156 debtors with gross annual incomes between $6,000 and $30,000 and unsecured debts at least three times the gross income. These debtors reported a total of $12.67 million in unsecured debt. For most of these debtors, the interest and fees alone on their unsecured debt would be more than one-half of their total income.

      IRS EXPENSE ALLOWANCES: The IRS has developed a schedule of expense allowances for use in determining how much income a taxpayer has available to pay taxes that are in arrears. These IRS allowances were an integral part of the various forms of means testing proposed in 1998. The IRS schedule includes allowances for the four following general expense categories.

1. Housing: The housing allowance includes expenses for rent or mortgage payments, taxes and insurance, maintenance and repairs, homeowner fees, and utilities. It is determined by the county of residence and the size of the household. Three allowances are listed for each county -- for households of one or two persons, households of three persons, and households of four or more.

     A small proportion of the Over Median debtors in our sample reported owning either a second home or a rental property. It is not clear how mortgage payments and other costs of these properties would be treated under means testing.

2. Food: The IRS food allowance covers the cost of food, clothing, housekeeping supplies, personal care products and services. The amount is based on family size and gross family income. This can lead to circumstances in which a single person receives a higher monthly food allowance than a family of six with a much smaller gross income. Also, a family just above an income threshold would be treated as having less excess income than a family just below the threshold, all other factors being equal.

3. Transportation: This allowance covers the expense of owning and operating cars and using public transportation. The allowance is based on the number of cars owned, with a maximum of two, and the location described as either one of 26 metropolitan statistical areas or four regions in the country. The allowances vary from a low of $126 for an individual without a car in Buffalo, New York, to a high of $983 for an individual in Dallas, Texas who is making payments on two cars. Regardless of location, the allowance is generally about $700 per month higher for people making payments on two cars than for people who have no car.

4. Other Necessary Expenses: The expenses covered by this category include taxes, health care, court ordered payments, involuntary payroll deductions, secured debt payments, child and dependent care, life insurance, charitable contributions, educational costs, union and professional dues, and other miscellany.

     We do not know whether the bankruptcy courts will apply this or a similar criterion for allowing expenses in proposed chapter 13 plans. The situation seems likely to promote considerable litigation. As local standards evolve for each expense in this category, more debtors are likely to claim the maximum allowable amount on their monthly expenses.

     Further, application of the IRS expense guidelines for means testing will allow for increased pre-bankruptcy planning by debtors. For example, some debtors could reduce overtime or quit second jobs to reduce income to fall below the national median level. Debtors could also increase expense claims in a variety of ways. Application of the IRS allowances could allow a debtor to shelter an income several times the national median.

    We should note one detail about the means tests proposed in the Conference bill. Like earlier bills, the Conference bill sets a percent-of-unsecured debt repayable threshold for requiring the debtor to file in chapter 13. The particular percentage is 25%, to be paid over the 60-month duration of the plan. Unlike earlier bills, the Conference bill supplements the 25% test with an alternative minimum repayment of $5,000, or $83.33 per month. A debtor who met either of these tests would be required to file in chapter 13. This second test is a sound addition to the legislation because it prevents wealthy debtors with extremely high levels of unsecured debt from escaping chapter 13 by virtue of a percent-of-debt test alone. Several of the wealthiest debtors in our sample would not have been subject to chapter 13 without imposition of the second test.

     When all of these factors are considered, we believe that the final return to unsecured creditors under means testing as proposed would be less than $1 billion annually. This figure is in agreement with the results reported by Culhane and White.⁽¹³⁾

    UNLIMITED HOMESTEAD EXEMPTIONS: One of the most controversial elements in consumer bankruptcy involves the unlimited homestead exemptions allowed in Florida, Texas, Kansas, Iowa, and South Dakota. There is concern that some debtors can discharge their debts in chapter 7 and emerge from bankruptcy relatively wealthy. Several particularly egregious examples of this have frequently been cited in newspaper articles and Congressional testimony.

CONCLUSIONS:

     1. Only a small percentage of current chapter 7 debtors have income sufficient to repay any portion of their unsecured debts.

     2. The means tests contained in the Conference bill would result in less than $1 billion annually being returned to unsecured creditors. This is much less than the $4 billion to $5 billion estimates published in some other reports.

     3. Using IRS guidelines as expense allowances will be cumbersome and conducive to "gaming" the system and adding to bankruptcy litigation.

     4. Concerns about debtors abusing large homestead exemptions in some states were not validated in our sample. Such cases must be very rare.

    5. Using a means test that specifies a minimum threshold percentage of unsecured debt that must be payable in chapter 13 is not desirable, because it will exempt from chapter 13 many debtors who have the largest repayment capacities.

APPENDIX 1
NATIONAL DISTRIBUTION OF THE STUDY SAMPLE

End Notes:

1. Barron, J., and M. Staten, Personal Bankruptcy: A Report on Petitioners' Ability to Pay. Credit Research Center, Georgetown University, October, 1997. See also Personal Bankruptcy: The Credit Research Center Report on Debtors' Ability to Pay. General Accounting Office, Report GAO/GGD-98-47, February, 1998.

2. Neubig, T., and F. Scheuren, Chapter 7 Bankruptcy Petitioners' Ability to Repay: the National Perspective, 1997. Ernst & Young L.L.P., March, 1998.

3. Culhane, M., and White, M., Means-Testing for Chapter 7 Debtors: Repayment Capacity Untapped? American Bankruptcy Institute, November, 1998.

4. Our decision to include only no-asset cases in the sample was based on three considerations. First, the number of original chapter 7 consumer cases in which assets are eventually distributed is approximately 1% of a year's filings. Asset cases also remain pending much longer than no-asset cases. Sampling to obtain a reasonable number of asset cases therefore requires a different procedure from the one used here to obtain a proportional sample of manageable size. Second, the geographical distribution of asset cases may be different from the distribution of no-asset cases. Hence, asset cases should be studied separately. And third, asset cases, by definition, generate some repayments to creditors, and the intent of means testing is primarily to capture repayments from debtors who now pay nothing. We emphasize, nevertheless, that it would be valuable, in a separate study, to compare the repayments by debtors in asset chapter 7 cases to potential repayments by those debtors under chapter 13 plans.

5. Of this total 23,131 (2.37%) were filed in North Carolina and Alabama, which are served by Bankruptcy Administrators rather than United States Trustees.

6. House and Senate bills used different measures of national median income throughout the 105^th Congress. We have described the consequences of these differences in Flynn, E. and G. Bermant, Measuring Means-Testing: It's All in the Words, American Bankruptcy Institute Journal 1 (September, 1998.) The Conference bill contained both sets of standards at different places: the House standards to amend 11 U.S.C.�704, and the Senate standards to amend 11.U.S.C.�707.

7. We have used arithmetic averages (means) here because the population has been divided in advance into separate income groups.

8. Some discretionary judgment is required for this analysis, but we are confident that the results would be essentially the same if the work had been done by other competent analysts.

9. We arrive at this national estimate by multiplying the amount available in our sample by 500. See our sampling method described above. Each annual cohort of debtors would provide one-fifth of this amount each year; after five years of operation, the total amount would be realized each year.

10. Discussions of the effects of stigma as a factor in bankruptcy decisions may be found in Fay, S., E. Hurst, and M. White, The Bankruptcy Decision: Does Stigma Matter? Unpublished ms., Department of Economics, Univ. Michigan (1998); Gross, D., and N. Souleles, Explaining the Increase in Bankruptcy and Delinquency: Stigma versus Risk-Composition. Unpublished ms., Graduate School of Business, Univ. Chicago and the Wharton School, Univ. Pennsylvania (1998).

11. Of course the debtor would have to be prudent enough to acquire the cars early enough to avoid the window of time before filing during which the purchases would be disallowed.

12. Internal Revenue Service, Handbook 105.1, Collecting Contact Hankbook, Chapter 3: Analyzing Financial Information (9/26/96)

13. See supra note 3.