Did bankruptcy reform fail? An empirical study of consumer debtors

American Bankruptcy Law Journal

Volumn 82, Issue 3, 2008, Pages 349-406

  Lawless, Robert M ^a   Littwin, Angela K ^b   Porter, Katherine M ^c   Pottow, John A E ^d   Thorne, Deborah K ^e   Warren, Elizabeth ^f  

^a UNIVERSITY OF ILLINOIS AT URBANA CHAMPAIGN   (United States)

^b UNIVERSITY OF TEXAS   (United States)

^c UNIVERSITY OF IOWA COLLEGE OF LAW   (United States)

^d UNIVERSITY OF MICHIGAN LAW SCHOOL   (United States)

^e OHIO UNIVERSITY   (United States)

^f HARVARD LAW SCHOOL   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 57149096512     PISSN: 00279048     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (217)

1. Median family income peaked in 2000 at $61,049 and has not yet returned to that level. U.S. CENSUS BUREAU, MONEY INCOME OF FAMILIES-PERCENT DISTRIBUTION BY INCOME LEVEL, RACE, AND HISPANIC ORIGIN, Table 673, available at http://www.census.gov/compendia/ statab/tables/08s0673.xls (last visited Sept. 3, 2008) (constant 2007 dollars).
2. See Bureau of Labor Statistics Consumer Expenditure Survey (reporting survey data's finding of consumer expenditure increases in 2005 and 2006), available at http://www.bls.gov/news.release/cesan.nr0.htm (press release announcing completion of data).
3. By the fourth quarter of 2007, total outstanding consumer debt was approximately $2557 billion. Federal Reserve Statistical Release, Consumer Credit, G.19, available at http://www.federalreserve.gov/ releases/gl9/Current. In 2001 total outstanding consumer credit was $1963 billion (in 2007 dollars).
4. see id. at Release Date Jan. 8, 2007, available at http://www.federalreserve.gov/releases/gl9/20070108, which means there was an increase of $594 billion in the six-year time span.
5. The Federal Reserve Board reports a 4.70% credit card charge-off rate for the first quarter of 2008, an increase of 18% from a year earlier. Federal Reserve Statistical Release, Charge-Off and Delinquency Rates on Loans and Leases at Commercial Banks, available at http://www.federalreserve.gov/ releases/chargeoff/chgallnsa.htm (last visited Sept. 3, 2008).
6. See Jenn Abelson, Entering the Repossession Lane: Default Rate Soars on Auto Loans in Pattern Likened to Mortgage Crisis, THE BOSTON GLOBE, Mar. 7, 2008, at 1A (reporting increase in regional and national auto foreclosures).
7. The Mortgage Bankers Association reported that the percentage of loans in the foreclosure process was 2.47% at the end of the first quarter of 2008. The percentage of loans on which foreclosure actions were started during the quarter was 0.99% on a seasonally-adjusted basis. Both the percentage of loans in foreclosure and the percentage of loans on which foreclosure actions were started were the highest since 1979. Press Release, Mortgage Bankers Association, Mortgage Bankers Association National Delinquency Survey (June 5, 2008), available at http://www.mbaa.org/NewsandMedia/PressCenter/62936. htm;
8. see also Vikas Bajaj & Michael M. Grynbaum, A Ruing Tide of Mortgage Defaults, Not All on Risky Loans, N.Y TIMES, June 6, 2008, at Cl (reporting data from the Mortgage Bankers Association);
9. Robert B. Avery, Kenneth P. Brevoort & Glenn B. Canner, The 2006 HMDA Data, FEDERAL RESERVE BULLETIN, Dec. 2007, at A102, available at http://www.federalreserve.gov/pubs/bulletin/2007/ pdf/hmda06final. pdf.
10. Growing Rich-Poor Divide in Affording Necessities: Economic Discontent Deepens as Inflation Concerns Rise, THE PEW RESEARCH CENTER FOR THE PEOPLE AND THE PRESS, Feb. 14, 2008, at 3, available at http://people-press. org/reports/pdf/395.pdf. Seventy-one percent of households making less than $50,000 annually reported that they were falling behind.
11. Bankruptcy Abuse Prevention and Consumer Protection Act of 2005, Pub. L. No. 109-8, 119 Stat. 23 (2005). There were 1.6 million filings in the twelve months that ended on June 30, 2005, the last annual period unaffected by the implementation of BAPCPA. Administrative Office of the U.S. Courts, Bankruptcy Statistics, available at http://www.uscourts.gov/bnkrpctystats/ statistics.htm#quarterly (last visited Sept. 3, 2008) (follow "2005 by Chapter" hyperlink under "12-month period ending June"). Because our sample was generated using data from the Automated Access to Court Electronic Records ("AACER"), all 2006 and 2007 national bankruptcy filing statistics also are from AACER. Filing data before 2006 are not available from AACER. Because data from the Administrative Office of the U.S. Courts ("AO") include transferred and reopened cases as new filings and the AACER data do not, the AO data may overstate bankruptcy filings by 2 to 3%.
12. See Posting of Robert M. Lawless to Credit Slips, Why 1 Use AACER Filing Statistics, http://www.creditslips.org/creditslips/2008/05/why- i-use-aacer.html (May 13, 2008).
13. Administrative Office of the U.S. Courts reports that there were 667,431 bankruptcy filings in the final quarter of 2005 (630,228 of which were filed in October 2005 - before BAPCPA went into effect). Contrast this to the 371,668 filings reported in the fourth quarter of 2004. Press Release, Administrative Office of the U.S. Courts, Bankruptcy Filings Surge in Calendar Year 2005 (Mar. 24, 2006), available at http://www.uscourts.gov/Press-Releases/ bankruptcyfilings032406.html.
14. See supra note 8.
15. "We have seen bankruptcy rates fall dramatically from about 2 million bankruptcies in 2005 to the point where I doubt there will be over 1 million bankruptcies in 2006, if current trends continue ... [F]or now, almost one year later, bankruptcy reform seems to have been a success." 152 CONG. REC. S10647-48 (daily ed. Sept. 29, 2006) (statement of Sen. Grassley).
16. See generally John Poirier, Personal Bankruptcy Cases Rise Despite Reforms, REUTERS, June 12, 2006 ("Some people think that merely reducing the number of filings regardless of who they are and what kinds of problems they have is a success. . . .") (quoting Professor Melissa Jacoby) (internal quotation marks omitted).
17. CONG. REC. 8509 (1999) (statement of Rep. Gekas).
18. See also H.R. REP. No. 109-031 (I), at 92 (2005), which summarized that a "factor motivating comprehensive reform is that the present bankruptcy system has loopholes and incentives that allow and - sometimes - even encourage opportunistic personal filings and abuse. . . [S]ome bankruptcy debtors are able to repay a significant portion of their debts, according to several studies. (airrent law, however, has no clear mandate requiring these debtors to repay their debts."
19. President George W. Bush, Remarks at the Signing of Bankruptcy Abuse Prevention and Consumer Protection Act (Apr. 20, 2005), available at http://www.whitehouse.gov/news/releases/2005/04/2005/0420-05.html ("[T]oo many people have abused the bankruptcy laws. They've walked away from debts even when they had the ability to repay them. . . . Under the new law, Americans who have the ability to pay will be required to pay back at least a portion of their debts.").
20. Hanover Nat. Bank v Moyses, 186 U.S. 181, 192 (1902).
21. Holden Lewis, What Bankruptcy Reform Means to Consumers, BANKRATE.COM, Mar. 16, 2001, http://www.bankrate.com/brm/news/pf/ 20010316a.asp (quoting Lynne Strang, spokeswoman for the American Financial Services Association).
22. CONG. REC. S1856 (daily ed. Mar. 1, 2005) (statement of Sen. Grassley).
23. See 11 U.S.C. § 707(b) (2006).
24. CONG. REC. S1779 (daily ed. Feb. 28, 2005) (remarks of Sen. Specter).
25. Ronald J. Mann, Consumer Bankruptcy and Credit in the Wake of the 2005 Act: Bankruptcy Reform and the "Sweat Box" of Credit Card Debt, 2007 U. ILL. L. Rev. 375, 398 fig.2.
26. For more details on the Consumer Bankruptcy Project, see infra Appendix I.
27. In order to arrive at nationally representative estimates, we weighted the data to adjust for the slight underrepresentation of respondents who filed under Chapter 13. For more details, see infra Appendix I, Part C.
28. See AACER, http://AACER.com (last visited Sept. 3, 2008).
29. We obtained a PACER fee waiver in every district except the Southern District of Texas, a consolidated district m which the district court, not the bankruptcy court, decides PACER fee waivers.
30. Administrative Office of the U.S. Courts: PACER Service Center, http://pacer.psc.uscourts.gov (last visited Sept. 3, 2008).
31. See Posting of Robert M. Lawless to Credit Slips, Monthly Filings. Jan. 2006-Mar. 2008, hup://www.creditshps.org/creditslips/2008/04/ monthly-filings.html (Apr. 17, 2008). The three exceptions were April, May, and December. Because the number of business days per month varies, we used the daily-averages for each month.
32. See id.
33. The questionnaire is reproduced in Appendix II, infra.
34. We detected no economically meaningful or systematic response bias. Non-respondents were similar to respondents with regard to income, assets, debts, net worth, market value of homes, and prior bankruptcy history. For more details, see infra Appendix I, Part B(3).
35. TERESA A. SULLIVAN, ELIZABETH WARREN & JAY LAWRENCE WESTBROOK, AS WE FORGIVE OUR DEBTORS: BANKRUPTCY AND CONSUMER CREDIT AND BANKRUPTCY IN AMERICA (Oxford University Press 1989) [hereinafter AS WE FORGIVE].
36. TERESA A. SULLIVAN, ELIZABETH WARREN & JAY LAWRENCE WESTBROOK, THE FRAGILE MIDDLE CLASS: AMERICANS IN DEBT (Yale University Press 2000) [hereinafter FRAGILE MIDDLE CLASS].
37. ELIZABETH WARREN & AMELIA WARREN TYAGI, THE TWO-INCOME TRAP: WHY MIDDLE-CLASS MOTHERS AND FATHERS ARE GOING BROKE (Basic 2003) [hereinafter TWO-INCOME TRAP].
38. Our prior studies were not random national samples but random samples drawn from five judicial districts. We have no reason to believe this affects the comparisons we make to these earlier cohorts of bankruptcy filers.
39. We have no reason to suspect a bias in the composition of this subset. Indeed, 86.7% of respondents indicated that they were willing to participate in a follow-up telephone interview See infra Appendix I, Part B(4) (discussing statistical testing suggesting lack of response bias) Teresa A. Sullivan, Elizabeth Warren & Jay Lawrence Westbrook, Folklore and Facts: A Preliminary Report from The Consumer Bankruptcy Project, 60 AM. BANKR. L.J. 293 (1986) [hereinafter Folklore and Facts] (1981 CBP data);
40. Teresa A. Sullivan, Elizabeth Warren & Jay Lawrence Westbrook, Consumer Debtors 'Ten Years Later: A Financial Comparison of Consumer Bankrupts 1981-91, 68 AM. BANKR. L.J. 121 (1994) [hereinafter Ten Tears Later] (1991 CBP data);
41. Teresa A. Sullivan, Elizabeth Warren & Jay Lawrence Westbrook, Less Stigma or More Financial Distress: An Empirical Analysis of the Extraordinary Increase in Bankruptcy Filings, 59 STAN L. REV. 213 (2006) [hereinafter Less Stigma] (2001 CBP data).
42. CONG. REG. S1856 (daily ed. Mar. 1, 2005) (statement of Sen. Grassley); accord Press Release, Senator Charles Grassley, Grassley Renews Effort to Reform Bankruptcy Code (Feb. 2, 2005), available at http://grassley.senate.gov/news/Article.cfm?customel/-dataPageID-l502=3544 ("The bankruptcy system currently allows anyone to get full debt cancellation in Chapter 7 with no questions asked, even if they have the means to pay off their debts.");
43. see also 151 CONG. REC. S1843 (daily ed. Mar. 1, 2005) (statement of Sen. Hatch) ("As matter of public policy we must say that those relatively high-income debtors, those capable of paying back their substantial debts, should at least pay something back."); 151 CONG. REC. S1813 (daily ed. Mat. 1, 2005) (statement of Sen. Frist).
44. See 151 CONG. REC. S1856 (daily ed. Mar. 1, 2005) (statement of Sen. Grassley) (proposing amendments to 11 U.S.C. § 707(b) (2004)).
45. Id.
46. Id.
47. U.S.C. §§ 109(e), 707(b) (2006).
48. See, e.g., 151 CONG. REG. S1813-6, S1856-57 (daily ed. Mar. 1, 2005) (statements of Sens. Frrst and Grassley).
49. SEE 151 CONG. REG. S1779 (daily ed. Feb. 28, 2005) (statement of Sen. Specter).
50. We should articulate a necessary assumption: that there was not a trend of rising income among those who tend to file for bankruptcy. If such a hypothetical trend were true, then merely showing similar incomes between bankruptcy filers before and after BAPCPA would suggest that BAPCPA succeeded in driving out the even higher-income debtors that we would have expected to see in bankruptcy pursuant to this trend. In light of the flat incomes of the U.S. population generally, see supra note 1, and our prior research suggesting that bankruptcy filers generally have lower (and stably lower) incomes than the general population,
51. see Less Stigma, supra note 34, at 222, we would find it difficult to construct a model predicting such a trend.
52. See 11 U.S.C. § 109(e) (2006).
53. We also recognize another assumption in this methodology: that the number of high-income abusers in the system was sufficient to have an effect on median income (and distribution). If Congress was trying to target only one or two bad apples, then presumably their exclusion by a successful means test would be difficult to detect in looking at the median incomes of a random national sample of the hundreds of thousands of bankruptcy filers. If, by contrast, Congress were going after a scourge of deadbeats, then we would expect to see some change in income patterns. We think that while some members of Congress tried to soothe that the means test wouldn't affect that many people, see notes 12 and 16 and accompanying text, they insisted at the same time that there was a "bankruptcy crisis" of debtors who could pay their debts. 151 CONG. REC. S1855-56 (daily ed. Mar. 1, 2005) (statement of Sen. Grassley). Accordingly, we take these advocates at their word that the nation was facing an epidemic "bankruptcy crisis" that had "contributed to the fraying of the moral fiber of the nation." Id. at S1856. (We incidentally note that in our analysis of the distribution of incomes among the 2007 filers,
54. see infra Figure 2, we find similar distributions of incomes among the subset of debtors with above-median incomes, even those with median household incomes above $60,000, which suggests little change in income profile amongst the highest income echelons of filers.)
55. The 1981-2007 difference is significant (p < .0001). The 1991-2007 difference is not significant (p < .92). The 2001-2007 difference is not significant (p < .86). For more details, see infra Appendix III.
56. The 1991 income figure differs significantly from neither the 2001 nor the 2007 income figures, and the 2001 income figure does not differ significantly from the 2007 figure. In effect, bankrupt debtors had the same income levels in 1991, 2001, and 2007. For 2007 compared with 2001, Wilcoxon Z = 2.131, p = .0331; t-test = .175, p = .860. For 2007 compared with 1991, Wilcoxon Z = 1.387, p = .165; t-test = .100, p = .920.
57. The 1981 income figure is significantly greater than the incomes of the debtors in 1991, 2001, and 2007 (p < 0.0001). For precise numbers, see Appendix III.
58. See also infra note 52 (regarding debtor income data recently released by the Administrative Office of the U.S. Courts).
59. See supra note 46 (2007 compared with 2001).
60. There were 465,878 Chapter 13 cases of 1,618,062 non-business cases filed in fiscal year 2004 compared with 276,649 Chapter 13 cases of 673,615 non-business cases filed in fiscal year 2007. See Press Release, Administrative Office of the U.S. Courts, Bankruptcy Filings Drop 61 Percent in March 2007 12-Month Period (June 27, 2007), available at http://www.uscourts.gov/ Press-Releases/bankruptcyfil-ings062707.html. The fiscal year for these data ends March 31.
61. Whether Chapter 13 filers have the same income as their Chapter 7 counterparts is technically distinct from the dividend to unsecured creditors those Chapter 13 filers pay. One new study (drawn only from the Tenth Circuit) finds that Chapter 13 filers proposed to pay higher dividends to unsecured creditors in their plans in the year immediately following BAPCPA than in the year immediately before.
62. See Bruce M. Price & Terry Dalton, From Downhill to Slalom: An Empirical Analysis of the Effectiveness of BAPCPA (and Some Unintended Consequences), 26 YALE L. & POLY REV. 135, 193-94 (2007). It will be interesting to see whether these proposals translate into a change in actual dividends. If anything, what we know about actual dividends to unsecured creditors in Chapter 13 is discouraging: the median payment is $0.
63. See Scott F. Norberg, The Chapter 13 Project: Little Paid to Unsecureds, 26 AM. BANKR. INST. L.J. 1, 54-56 (2007);
64. Scott F. Norberg & Andrew J. Velkey, Debtor Discharge and Creditor Repayment in Chapter 13, 39 CREIGHTON L. REV. 473 (2006) .
65. Wilcoxon Z = 1.718, p = .086.
66. Wilcoxon Z =.894, p = .371. Our data on income in Chapter 7 and 13 are consistent with, but somewhat lower than, recently released data from the Administrative Office for U.S. Courts ("AO"). See ADMINISTRATIVE OFFICE OF THE U.S. COURTS, 2007 REPORT OF STATISTICS REQUIRED BY THE BANK-RUPTCY ABUSE PREVENTION AND CONSUMER PROTECTION ACT OF 2005, available at http://www.uscourts.gov/bnkrptcystats/2007/BAPCPAstats.html (last visited Sept. 3, 2008) (first annual report on bankruptcy statistics statutorily required by BAPCPA). The AO data are interesting but difficult to compare with the CBP data because of the AO's reporting conventions, some of which were dictated by the statutory mandate.
67. See generally 28 U.S.C. § 159 (2006). For example, because it was required to report on all cases filed in 2007, the AO report includes two related (and apparently abusive) mean-skewing Chapter 13 cases that listed almost $2.4 billion in assets. Id. at 6 n.4. Removing these two outlier cases alone drops the mean value of total assets from $175,234 to $160,187. Without knowing the underlying distribution of the data in the AO report, it is impossible to make meaningful comparisons of means between the variables reported in it and the CBP study. Nevertheless, one variable for which the AO reports median statistics as does the CBP is debtor income. The AO finds median monthly income of $2,150 and $3,146 in Chapter 7 and 13 cases respectively. These data, however, exclude cases where the debtor reported debts "predominately of a business nature" and where "any information on the schedules was incomplete." Id. at tbl. 1A n.l (reporting general caveat on data collection procedures). Selecting similar cases from the CBP database produces roughly comparable median monthly income figures of $1,926 and $3,033 for Chapter 7 and 13 cases respectively. While these numbers appear statistically significantly lower than the AO numbers, the absolute dollar difference is not large. More importantly, this comparison is at best a guess regarding what algorithms the AO used to exclude "incomplete" data. For example, our hand-coding allowed us to record barely legible handwritten petitions from pro se debtors (who by hypothesis might have lower incomes), but the AO's dataset may exclude these figures. We cannot know. Thus the mild apparent divergence between the AO data and CBP data remain untroubling given the necessary imperfection of coding "incomplete" bankruptcy records. More importantly, we note that our approach biased our results against our conclusions; if we were to use the AO data of slightly higher debtor incomes, our convictions of BAPCPA's inefficacy would only strengthen.
68. While most members of Congress charged ahead, some academies worried that BAPCPA would have just this perverse outcome. In a joint letter to Congress (signed by some ot us), about one hundred law professors explained: Our problem is not with means-testing per se. Our problem is with the collateral costs that this particular means-test would impose. This is not a typical means test, which acts as a gatekeeper to the system. It would instead burden the system with needless heatings . . . and arbitrarily deprive families of needed relief. Letter from Bankruptcy and Commercial Law Professors to Senators Specter and Leahy (Feb. 16, 2005), available at http://www.abiworld.org/pdfs/ LawProfsLetter.pdf. Special prescience in this regard goes to Professor Jean Braucher, who wrote in a 2002 symposium right in the midst of the debates: The current debate is not really about whether to means-test bankruptcy, but about whether it is possible to do a better job of catching the system's abusers without excluding many "honest but unfortunate" debtors in the process. . . . With only modest gams possible in catching abusers and increasing collections in bankruptcy, it is particularly appropriate to ask whether any given proposal for reform will do more harm than good, imposing new costs for minimal returns. We should be asking: would the proposed changes make for a better system than we currently have? The pending consumer bankruptcy legislation fails this test. It would make access to bankruptcy more difficult for all, imposing new costs and hurdles and thus pricing the worst off out of the system.
69. Jean Braucher, Means Testing Consumer Bankruptcy: The Problem of Means, 7 FORDHAM J. CORP. & FIN. L. 407, 408-09 (2002).
70. See supra note 45.
71. U.S. CENSUS BUREAU, CURRENT POPULATION REPORTS: CONSUMER INCOME, INCOME, POVERTY, AND HEALTH INSURANCE COVERAGE IN THE UNITED STATES: 2006, P60-233, at tbl.l, available at http://www.census. gov/prod/2007pubs/p60-233.pdf. With inflation adjustment, that would have been $49,573 in 2007.
72. See Less Stigma, supra note 34, at 222.
73. Philip Shenon, Home Exemptions Snag Bankruptcy Bill, N.Y. TIMES, Apr. 6, 2001, at A1.
74. See, e.g., 148 CONC. REC. S5087 (2002) (statement of Sen. Kohl) (announcing proposal to cap homestead equity at $175,000).
75. The other states that place no limit on the value of the exempt homestead are Iowa, Kansas, Oklahoma, and South Dakota, as well as the District of Columbia. COLLER ON BANKRUPTCY (15th ed., rev. vol.);
76. see also David A. Samole & David L. Rosendorf, Homestead Exemption No Longer "Debtors' Paradise", 24-10 AM. BANKR. INST. L.J. 6 (Dec. 2005-Jan. 2006).
77. See, e.g., 11 U.S.C. § 522(p) (2006) (restricting under certain circumstances homestead exemptions to $125,000 in equity).
78. For 2007 compared with 2001, Wilcoxon Z = -2.293, p = .0218. For 2007 compared with 1991, Wilcoxon Z = -8.437, p < .0001. From 2007 to 2001, mean asset value increased from $80,787 to $100,878 (t-test =4.63, p < .0001).
79. For 2007 compared with 2001, Wilcoxon Z = -2.293, p = .0218. For 2007 compared with 1991, Wilcoxon Z = -11.844, p < .0001. For 2007 compared with 1981, Wilcoxon Z = -14.892, p < .0001.
80. Since 1981, the proportion of families filing for bankruptcy who own their own homes has remained fairly constant at about half of all debtors. About 51.6% of the filers in 2007 lived in their own homes, a difference that is not statistically significant from the 54.1% who did in 2001 (p = .165). In 1991, about half of all the non-Philadelphia debtors were homeowners, although only about 43.9% of the total sample were homeowners. FRAGILE MIDDLE CLASS, supra note 30, at 202. (The authors believe that the higher proportion of legal service clients in Philadelphia drove down the average homeownership rates in that district unrepresentatively. See Ten Years Later, supra note 35, at 132.) In 1981, 52% of the debtors were homeowners. As WE FORGIVE, supra note 29, at 129.
81. Wilcoxon Z = -1.314, p = .189.
82. t-test = 4.62, p < .0001.
83. Wilcoxon Z = -2.230, p = .026.
84. See National Association of Realtors®, Median Sales Price of Existing Single-Family Homes for Metropolitan Areas, available at http://www.realtor.org/Research.nsf/files/msapricesf.pdf (last visited Sept. 3, 2008) (2007 data); Real Estate ABC, USA Real Estate Median Sales Price of Existing Homes since 1968, available at http://www.realestateabc.com/ graphs/natlmedian.htm (last visited Sept. 3, 2008) (archiving National Association of Realtors historical data on median home values) (2001 data).
85. See U.S. CENSUS BUREAU, AMERICAN HOUSING SURVEY, available at http://wwvv.census.gov/ hhes/www/housing/ahs/ahs05/tab3- 15.pdf (last visited Sept. 3, 2008). Although per-household data are not yet available on 2007 mortgage loads, aggregate data are available. The Federal Reserve Board reports that total aggregate household mortgages outstanding were $10,533 trillion in 2007. See Federal Reserve Statistical Release, Flow of Funds Accounts of the United States, Table Z.1, available at http://www. federalreservc.gov/rcleases/z1 http://federalreservc.gov/rcleases/z1 (last visited Sept. 3, 2008). With interest rates on adjustable-rate mortgages only increasing since 2007, and the housing crash unfolding, we have strong reason to suspect that both housing debt and property values will worsen for consumers.
86. Wilcoxon Z = -7.9629, p < .0001.
87. Mean (median) non-home assets in 2001 were $24,229 ($11,123) and in 2007 were $27,488 ($13,227). The difference in means was almost statistically significant (t = 1.73, p = 0.084), while a Wilcoxon rank-sum test showed a statistically significant difference between the distributions (z = 2.21, p = 0.027).
88. There is not a perfect correspondence between non-mortgage debt and non-home assets. Tax liens, for example, are non-mortgage debt that can encumber both home and non-home assets. Accordingly, our data cannot be read to assume that the median debtor has encumbered his non-home assets by 300%.
89. See Sumit Agarwal, John C. Driscoll, Xavier Gabaix & David Laibson, Learning in the Credit Card Market (Nt'l Bureau of Econ. Research, Working Paper No. W13822, Feb. 2008), available at http://www.nber.org/ papers/wl3822.pdf http://www.nber.org/papers/w13822.pdf (last visited Sept. 3, 2008), at Appendix A, which reports a panel dataset that contains three years of credit card statements, representing 120,000 consumers and 4,000,000 credit card statements. The issuer of the analyzed credit cards in 2002-04 was charging average annual late/overlimit/cash advance fees of $141 per account and $226 in penalty interest. These charges were on accounts that showed a mean monthly balance of $1735 (with a standard deviation of $1978). This surcharge was roughly 21% above the base APR. For a set of 120,000 cardholders, the issuer would net $44 million per year from these fees and penalty interest charges.
90. Total secured debt grew at a 3.5% average annual, non-compounded rate. In the same time period, unsecured debt listed in bankruptcy, which had grown at about 3% per year from 1991 to 2001, increased by about 7.3% annually from 2001 to 2007. By contrast, income stagnated at about $27,000 from 1991 forward.
91. See supra note 45 and accompanying text; infra Appendix III.
92. The exact medians, calculated on a debtor-by-debtor basis, using constant 2007 dollars, are reported in Appendix IV, infra, along with Wilcoxon Z and p values that show the differences between all years are statistically significant. Another study found similarly significant increases in the value ot debtors' real and personal property, as well as secured and total debt in Chapter 13. (They did not provide figures for Chapter 7 debtors or debt-to-asset ratios.) The authors caution, however, that the spike in filings immediately before BAPCPA took effect may have influenced their results.
93. See Price & Dalton, supra note 50, at 191-92.
94. Debt-to-income ratios are just one measure of financial distress. There are others. Consider the Debt Service Ratio (DSR) and the Financial Obligations Ratio (FOR), two measures of after-tax income available to pay minimum debts (DSR) and minimum debts plus residential leases and other regular payments (FOR). These ratios, compiled by the Federal Reserve Board, show upward trends since 1980, from about 16 to 18% and 10 to 14% respectively. See BOARD OF GOVERNORS OF THE FED. RESERVE SYS., REPORT TO THE CONGRESS ON PRACTICES OF THE CONSUMER CREDIT INDUSTRY IN SOLICITING AND EX-TENDING CREDIT AND THEIR EFFECTS ON CONSUMER DEBT AND INSOLVENCY 12-13 & fig.3 (2006). "One problem with these data, however, is that minimum payments for credit cards may result in negative amortization.
95. Another is that they are aggregate data and hence do not show heterogeneity; case study data suggest bankrupt debtors bear a disproportionately heavy debt service burden." John A. E. Pottow, Private Liability for Reckless Lending, 2007 U. ILL. L. REV. 407, 408-09 n.16.
96. The median total debt-to-income ratio for 1981 was 1.411. For 2007, the median total debt-to income ratio was 3.272. For 2007 compared with 1981, the difference is statistically significant (Wilcoxon Z = -22.531, p < .0001).
97. See generally Mann, Sweat Box, supra note 19, at 389-92.
98. The difference is statistically significant (Wilcoxon Z = -24.279, p < .0001).
99. FolKlore and Facts, supra note 34, at 320.
100. Sce 11 U.S.C. §§ 707(b)(6), 707(b)(7) (2006).
101. See supra notes 57-60 and accompanying text.
102. The only way to stretch BAPCPA to indicate some interest in net worth is to re-envision the credit counseling requirement as a way to screen "can-pay" debtors in an undifferentiated, unspecified way wholly distinct from the means test's income screen. Thus it could be, at least theoretically, that while the means test failed, credit counseling "worked" by inducing the stronger debtors (in terms of net worth and debt-to-income ratios) to soldier on and stay out of bankruptcy. While possible, this theory strikes us as deeply implausible. Credit counseling has been offered by independent providers who arc unlikely to coordinate an effort to herd families in the same direction. Indeed, credit counselors that work through lawyers, as do many of the largest services, have a positive incentive to move debtors directly into bankruptcy quickly and cheaply. If they failed to do so, their supply of debtors from referring lawyers would presumably dimmish. See National Association of Consumer Bankruptcy Attorneys website, which provides links/ advertisements for two major credit counseling firms on its "Tools of the Trade" webpage: http://www. nacba.org/tools.php (last visited Sept. 3, 2008). Moreover, credit counseling is perceived, and may well be, a largely pro forma technicality with little substantive effect. See Debt Relief Lowdown, Is Credit Counseling a Waste of Time?, Mar. 1, 2006 (describing National Association of Consumer Bankruptcy Attorneys study finding only 3% of 61,000 surveyed debtors had enough financial cushion that pre-bankruptcy counseling would have helped avoid bankruptcy), available at http://www.debtrelieflowdown.com/credit-counseling/index.html. We also note that the credit counseling services often provide the mandatory debtor education courses within bankruptcy, see Brett Weiss, Why 1 Hale Pre-Bankjuptcy Credit Counseling, BANKRUPTCY LAW NETWORK, Apr. 6, 2007, available at http://www.bankruptcylawnctwork.com/2007/04/06/why-i- hate'pre-bankruptcy-credit-counseling, which could create yet another incentive to move debtors quickly to a bankruptcy filing so that the customer can return for the mandatory course. Similarly, it is difficult to characterize specific amendments that apply across the board to all debtors (not just high-income ones) as designed to screen can-pay debtors from can't-pay ones.
103. See, e.g., 11 U.S.C. § 1325 (2006) (anti-cramdown expansion for Chapter 13 plans);
104. U.S.C. § 506(a)(2) (2006) (valuation rules for debtor collateral);
105. U.S.C. § 523 (2006) (expansion of non-dischargeability provisions). We see these provisions as designed to help specific constituencies of lobby-favored creditors. That they have an indirect effect on the debtor is necessarily true; that they were designed to ferret out "can-pay" debtors in the same way that the means test was expressly designed to do is not. Finally, we are unable to explain Congress's motivation in passing random punitive provisions designed for all debtors, such as the expansion of discharge intervals.
106. See, e.g., 11. U.S.C. §§ 727(a)(8)-(9), 1328(f) (2006). Needless to say, we do not recall any arguments in the debates that the marginal debtor who "can pay" her debts is one who is attracted by the prospect of returning to bankruptcy in seven years but who will hunker down and pay her creditors if she faces an eight-year bar.
107. Median incomes in the general population have declined slightly from 2005 to 2007- Total outstanding debt has increased, and both revolving and consumer debt have been on the rise from 2005 to 2007. See G.19, supra note 3. We think these data suggest a worsening (or at best a stagnating) financial position, not an improving one, in the general population.
108. We cannot analyse the variance of the general population data. Accordingly, we also cannot exclude the possibility that BAPCPA was sorting those with worsened debt loads from those with much worsened debt loads, with the latter being the "can't-pays" still in the system and the former the "can-pays" driven from it. As we mention below, see infra text accompanying notes 86-88, we do not recall rhetoric in the BAPCPA debates that Congress was planning to bar low-income people from filing bankruptcy who had debt loads even worse than the 2001 median debtor in bankruptcy.
109. It also requires a peculiar twisting of the bankruptcy filing data of steady year-by-year increases from 2001 through 2004: it would have to assume that the supply of low-income debtors had reached a cap and all subsequent filers were exclusively high-income.
110. See Less Stigma, supra note 34, at 229-31, 238-40.
111. We dismiss as implausible in light of the general population data the possibility that they had better financial circumstances than their 2001 counterparts.
112. See supra note 84.
113. Robert M. Lawless, The Paradox of Consumer Credit, 2007 U. ILL. L. REV. 347 (2007).
114. 152 CONG. REG. S10647-48 (daily ed. Sept. 29, 2006) (statement of Sen. Grassley).
115. Edith H. Jones & Todd J. Zywicki, It's Time for Means-Testing, 1999 BYU L. Rev. 177, 215-21.
116. Id. at 209 ("If discharge of debts is easy in bankruptcy, debtors will incur more debt. Conversely, if obtaining bankruptcy relief is difficult. debtors will be more reluctant to incur debts.");
117. see also id. at n.114
118. Id. at 208.
119. Professor Zywicki explains: Do consumers respond to incentives to file bankruptcy? The experience of the past few months strongly suggests "yes". Although this is obviously very casual empiricism, it is backed by a volume of economic theory that predicts that consumers would respond to anticipated changes in the bankruptcy laws exactly as they apparently have-by
120. We recognize that debts are accrued over months and years and that turning an ocean liner takes time. So while we express some hesitation at BAPCPA's prospects in reducing debt acquisition, we reserve final judgment to a later point in time after many more data have been gathered.
121. See, e.g., FRAGILE MIDDLE CLASS, supra note 30.
122. See id. at 155, 197; see also Li Gan & Tarun Sabarawal, A Simple Test of Adverse Events and Strategic Timing Theories of Consumer Bankruptcy (Nat'l Bureau of Econ. Research, Working Paper No. 11763, Nov. 2005), available at http://www.nber.org/papers/w11763.pdf (last visited Sept. 3, 2008) (constructing model using the Panel Survey of Income Dynamics data to conclude that personal bankruptcy filings are more likely determined by exogenous events than strategic behavior).
123. When bankruptcy filings decline sharply, the exogenous shock model is consistent with the continuing presence of families in need of bankruptcy. If they are indeed out there, as we suspect they are, then it likely means they are being excluded from bankruptcy inefficiently and unfairly.
124. RONALDJ.MANN, CHARGING AHEAD: THE GROWTH AND REGULATION OF PAYMENT CARD MARKETS 177 (2006).
125. See Lawless, supra note 89, passim.
126. TWO-INCOME TRAP, supra note 31, at 180.
127. Mann, supra note 19, at 378-79.
128. Id. at 392-97.
129. See, e.g., 11 U.S.C. § 521 (2006) (debtors' filing obligations).
130. See 11 U.S.C. §§ 109(h)(1), 727(a)(11), 1328(g) (2006).
131. See, e.g., 11 U.S.C. § 707(b)(4) (2006) (attorney liability sections).
132. One interesting new study conducted after BAPCPA finds not only that the change in law has been profitable for credit card lenders, but that none of that surplus has been shared with consumers. See Michael Simkovic, The Effect of 2005 Bankruptcy Reforms on Credit Card Industry Profits and Prices (Working Paper, July 8, 2008), available at http://ssrn.com/abstract= 1157l58 (finding consumers are paying higher interest rates, late fees and overlimit fees after the amendments' implementation despite the lobbying promise that BAPCPA would eliminate alleged "bankruptcy tax" on non-bankrupt borrowers).
133. See James J. White, Abuse Prevention 2005, 71 Mo. L. REV. 863 (2006).
134. Id. at 874.
135. Pearson chi-square = 37.570, p < .001; Cramer's v = .153, p < .001. It is significant that we can offer internal comparison with our 2001 data. For those who think our respondents self-congratulate themselves regarding how long they struggled before filing, we offer this comparative perspective: whatever the level of self-congratulatory distortion is (if any), there is no reason to assume it varies over time. Thus, the relative increase from 2001 to 2007 is noteworthy.
136. See Lawless, supra note 89, passim (presenting data on long-term correlation between debt and bankruptcy).
137. Of the 845 debtors who had been contacted by a debt collector, these were the recorded responses:
138. It was illegal to file bankruptcy: 4.9% (n=41)
139. You might go to jail if you filed: 3.9% (n=33)
140. I.R.S. would audit you if you filed: 7.3% (n=62)
141. You might lose your job if you filed: 8.5% (n=72)
142. Something else might happen if you filed: 19.4% (n=164) (text fields recorded).
143. Overall, 12.9% (n=133) reported one or more responses in categories 1 through 4. The total with a response in any of 1 through 5 was 23.6% (n=244).
144. Helen Huntley, Quiet in the Court, ST. PETERSBURG TIMES, Oct. 8, 2006, at ID (quoting Hon. Michael Williamson).
145. Jones & Zywicki, supra note 91, at 215-21.
146. "Bankruptcy has become so common that it has lost the stigma it had even a short generation ago. Today it is just another method lor getting out of debt, a tool just to get out of debt." 151 CONC. REC. S1813 (daily ed. Mar. 1, 2005) (statement of Sen. Frist).
147. See Less Stigma, supra note 34, at 238.
148. Id. at 239.
149. Sullivan, Warren & Westbrook also cite direct evidence on persisting stigma. See Less Stigma, supra note 34, at 243 (self-assessments of negative life events).
150. See id. at 246 (hiding bankruptcy from friends and family).
151. Deborah Thorne & Leon Anderson, Managing the Stigma of Personal Bankruptcy, 39 SOC. FOCUS 77, 77-97 (2006) (interviews with 37 newly bankrupt families).
152. Id. at 83.
153. See id. at 90.
154. l22 See Less Stigma, supra note 34, at 243.
155. "Personal bankruptcies are soaring because Americans have lost their sense of shame." Julie Kosterlitz, Over the Edge, 29 NAT'L J. 870, 871 (1997) (quoting then-Federal Reserve Board Chairman Alan Greenspan).
156. When the 1978 Code was enacted, express steps were taken to reduce the stigma of bankruptcy. The National Bankruptcy Review Commission Report observed, "The Code, for instance, replaced the term 'bankrupt' with 'debtor' and described a case filing as seeking an 'order for relief." EDITH H. JONES & JAMES I. SHEPARD, NAT'L BANKR. REV. COMM'N, ADDITIONAL DISSENT TO RECOMMENDATIONS FOR REFORM OF CONSUMER BANKRUPTCY LAW, available at http://govinfo.library.unt.edu/nbrc/report/24commvi08.html (last visited Sept. 3, 2008). We are unaware of any express efforts to cast more shame on and raise the sense of stigma of all debtors under BAPCPA.
157. See 11 U.S.C. § 707(b)(2)(A)(iii) (2006).
158. Unsecured debts are of course indirectly relevant in calculating the means test's threshold, see 11 U.S.C. § 707(b)(2)(A)(i)(I) (2006), but we spare our readers this tangent of statutory torture. So too we omit the means test's treatment of priority unsecured debts. See 11 U.S.C. § 707(b)(2)(A)(iv) (2006).
159. See Less Stigma, supra note 34, at 218 (describing the prior studies by year and numeral).
160. Some individuals may file for bankruptcy in either Chapter 11 or Chapter 12. For a discussion of this phenomenon, see Elizabeth Warren & Jay Lawrence Westbrook, Contracting Out of Bankruptcy: An Empirical Intervention, 118 HARV. L. REV. 1197, 1213 n.55 (2005);
161. Elizabeth Warren & Jay Lawrence Westbrook, Financial Characteristics of Businesses in Bankruptcy, 73 AM. BANKR. L.J. 499 (1999).
162. Robert M. Lawless & Elizabeth Warren, The Myth of the Disappearing Business Bankruptcy, 93 CAL. L. REV. 745 (2005) (reprinted in 47 CORP. PRAC. COMMENTATOR 593 (2005)).
163. Teresa A. Sullivan, Elizabeth Warren and Jay Lawrence Westbrook were the three principal investigators of CBP I.
164. Ten Tears Later, .supra note 34, at 125; see also AS WE FORGIVE, supra note 29 (describing the sampling method as a "systematic procedure that approximates random sampling").
165. Ten Years Later, supra note 34, at 150; Less Stigma, supra note 34, at 219.
166. Ten Years Later, supra note 34, at 125.
167. AS WE FORGIVE, supra note 29, at 344;
168. id. at 348 tbl.A2 (reporting sampling fractions).
169. Ten Years Later, supra note 34, at 125.
170. AS WE FORGIVE, supra note 29, at 63-69.
171. Less Stigma, supra note 34, at 218; AS WE FORGIVE, supra note 29, at 351-53.
172. Ten Years Later, supra note 34, at 125.
173. The same trio of tesearchers in CBP I were the principal investigators of CBP II.
174. Ten Years Later, supra note 34, at 125.
175. FRAGILF MIDDLE CLASS, supra note 30.
176. Id. at 266 (noting that no questionnaires were distributed in the Eastern District of Tennessee).
177. See id. at 268;
178. Less Stigma, supra note 34, at 219-20;
179. see also Ten Years Later, supra note 34, at 126 (describing the meeting of creditors that each debtor is required to attend).
180. See FRAGILE MIDDLE CLASS, supra note 30, at 263-64 (describing the 1991 questionnaire).
181. Less Stigma, supra note 34, at 220;
182. FRAGILE MIDDLE CLASS, supra note 30, at 277-79.
183. See Ten Years Later, supra note 34, at 126 (describing the total size of CPB II questionnaire sample).
184. See FRAGILE MIDDLE CLASS, supra note 30, at 266-68.
185. Id. at 269, 271.
186. Id. at 264;
187. see also Ten Tears Later, supra note 34, at 125-26.
188. FRAGILE MIDDLE CLASS, supra note 30, at 264.
189. Id. at 266;
190. see also Ten Years Later, supra note 34, at 126 (stating that "[f]ive cases were eliminated as outliers from the analysis of respondents").
191. FRAGILE MIDDLE CLASS, supra note 30, at 264.
192. TWO-INCOME TRAP, supra note 31, at 181-82.
193. Elizabeth Warren, Financial Collapse and Class Status: Who Goes Bankrupt?, 41 OSGOODE HALL L. REV. 115, 122 (2003).
194. Id. at 123 n.16.
195. Less Stigma, supra note 34, at 219.
196. TWO-INCOME TRAP, supra note 31, at 182-83.
197. A supplemental sample of additional debtors who were homeowners was also collected. This Article uses only the core sample of 1250 cases described in the text. TWO-INCOME TRAP, supra note 31, at 182.
198. See id. at 183-85 (explaining that some debtors completed the questionnaites and some did not).
199. A copy of the 2001 questionnaire is reprinted in Elizabeth Warren, Bankrupt Children, 86 MINN. L. REV. 1003 (2002).
200. TWO-INCOME TRAP, supra note 31, at 185.
201. Id. at 186.
202. Less Stigma, supra note 34, at 218.
203. TWO-INCOME TRAP, supra note 31, at 185.
204. Id. at 186.
205. The 2007 national filing data were supplied through the generous assistance of Mike Bickford and his colleagues at Automated Access to Court Electronic Records ("AACER"), an Oklahoma City-based bankruptcy data and management company. The questionnaire was translated into Spanish by a graduate student of Spanish. To improve the quality of the translation, a different graduate student translated the Spanish version back into English. This back-translation was compared with the English version of the questionnaire. Four petitioners who filed chapters other than 7 or 13 were incorrectly included in the sample. Thus, the total sample size is 5251 rather than 5255. The 5251 includes the original 5000 case sample plus the 275 cases selected as replacements for those debtors in the original sample to whom correspondence was returned as undeliverable. We used the data for non-business cases from the months of February 2007 and March 2007 and used only the numbers of Chapter 7 and Chapter 13 cases (excluding other chapters of relief) in calculating the percentages by chapter in the distribution. Administrative Office of the U.S. Courts, June 2007 Filings Drop When Compared to 2006 But Slow Upward Crawl Seen in Filings, Table F-2, available at http://www.uscourts.gov/bnkrpctystats/june2007/bankrupt-f2filmn.xls (last visited Sept. 3, 2008) (follow "Table F-2, 12-month" hyperlink).
206. This figure was provided courtesy of Mike Bickford and his colleagues at AACER. See supra note 167.
207. See note 28, supra, for more analysis of potential response bias.
208. For example, spouses were not to be coded as dependents. If a case had been incorrectly coded this way, the coding was changed to temove the spouse as a dependent. These cleaning criteria were applied to all cases. 258 cases were blind receded from the core sample, which is 10% of the respondent sample of 2521. The variables that categorize non-priority unsecured debts listed on Schedule F by type were excluded from the error check because by necessity and design they required coders to exercise some discretion in coding. These variables are Def Credit Card, Prob Credit Card, Student Loans, Rent, Utilities, and Medical. Debts that did not fit into one of these categories were not categorized but were included in the total unsecured debt.
209. For three variables relating to tallies of creditors, a greater margin of error was used to determine if an error existed. Many schedules contain debts listed "for notice only" or that duplicate prior debts yet are scheduled again to ensure a collection agency or law firm employed by the creditor is listed on the matrix of creditors. In some cases, it was not clear whether a debt was being scheduled for one of these quasi-superfluous purposes. Thus, some discretion had to be exercised in determining the number of different creditors. For the variables Tally Priority E and Tally Sec Debt, a recheck of the original coding was performed only if the original coding and recoding differed by thtee or more. For the variable Tally Unsec F, a recheck of the original coding was performed only if the original coding and recoding differed by five or more. If the difference between the original coding and the recoding were within these margins, no rechecking was performed.
210. The decision to interview 1032 respondents was based on a goal of conducting at least 1000 interviews. The interviewers were given lists of names from the 2007 volunteers from which they tried to make contacty by telephone, email or letter to schedule an interview. If no contact could be made within three weeks (leaving a maximum of three messages), the volunteer was removed from the list and replaced with another. Coincidentally, by the time the goal of 1000 interviews had been achieved, the interview team had exhausted the pool of volunteers. To test for possible response bias, the group of volunteers who completed interviews were compared with the group of volunteers who could not be be tracked down and thus did not complete interviews. No significant differences emerged regarding filing status, filing chaptet, total assets, total debts, total priority debts, monthly income, home value or other variables.
211. All of the interviewers were women. They ranged in age from their mid-20s to late-50s. The lead interviewer, Denise McDaniel, has worked on several prior empirical studies of bankruptcy, including interviewing debtors as part of CBP III.
212. For example, some respondents scheduled interviews but could not be reached at the arranged time or had phones that were disconnected in the interim, which required interviewers to make repeated efforts to complete the interview. Many debtors had irregular work schedules that required interviewers to contact them very early in the morning or late in the evening. In some instances, the stories of many debtors were sufficiently distressful or emotionally draining that interviewers could complete only a few interviews before they experienced fatigue. Note also that a small number (n=69) of initial volunteers subsequently requested not to complete the telephone interview for reasons ranging from illness to disinclination to spend cellular phone airtime minutes.
213. We ran all analyses using the unweighted data as well. The comparative findings with the prior studies were identical using either the weighted or unweighted CBP IV sample.
214. The inflation-adjustment figure used was for the first half of 2007 from the Consumer Price Index-All Urban Consumers-Not Seasonally Adjusted (Series ID CUUR0000SA0). That figure was 205.709 with a base period of 1982-1984. See Bureau of Labor Statistics, http://www.bls.gov/cpi (last visited Sept. 3, 2008) (follow "Databases & Tables" hyperlink; then follow "Top Picks" hyperlink next to "All Urban Consumers (Current Series) (Consumer Price Index-CPI)").
215. Because the data were inflation-adjusted to 2007 dollars, the numbers reported for these samples do not match the numbers in prior published studies in which the data were inflation-adjusted to a different year. To check that our analyses were consistent with our past findings, we also ran the original data (without inflation-adjustment) for each year and compared it with prior published findings of such data.
216. The outlier criteria inflation-adjusted to 2007 dollars and applied to the CBP IV sample were: total assets greater than $1,129,756, total debt greater than $1,129,756, and income greater than "146,869.
217. From the 1981 sample, 1.8% (n=27) ot the cases were eliminated as outliers. From the 1991 sample, 0.9% (n=6) of the cases were eliminated as outliers. From the 2001 sample, 0.9% (n=l 1) of the cases were eliminated as outliers. From the 2007 sample, 0.7% (n=18) of the cases were eliminated as outliers