Citation: 43 Rutgers L. Rev. 185 1990-1991
Content downloaded/printed from
Thu May 31 09:45:28 2012
-- Your use of this HeinOnline PDF indicates your acceptance
of HeinOnline's Terms and Conditions of the license
agreement available at http://heinonline.org/HOL/License
-- The search text of this PDF is generated from
uncorrected OCR text.
-- To obtain permission to use this article beyond the scope
of your HeinOnline license, please use:
SOCIAL SCIENCE RESEARCH ON
As We Forgive Our Debtors: Bankruptcy and
Consumer Credit in America.
By Teresa A. Sullivan, Elizabeth Warren & Jay L. West-
brook. New York: Oxford University Press 1989. Pp. 370.
Reviewed by Philip Shuchman*
P RE FACE ........................................... 186
I. W HAT W AS DONE ............................... 187
II. KNOWLEDGE AND LEGISLATION ................... 188
III. SOME OF THE AUTHORS' CLAIMS ................ 188
IV. PROBLEMS IN THEIR APPROACH TO EMPIRICAL RE-
SEA R CH .. .. ... .. .... .. .. .... .. .... .. ... ... .. 190
V. SITES ........................................... 192
VI. CRITICISMS OF THE ADMINISTRATIVE OFFICE STA-
T ISTIC S . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 19 7
VII. PROBLEMS IN THEIR METHODS ................... 198
A. 1981 Incomes of Chapter 7 Debtors ....... 198
B. Their Tenuous Inference ................ 199
VIII. INCOME CHANGES ................................ 201
IX. BUSINESS-RELATED PERSONAL BANKRUPTCIES .... 205
X. FEMALE DEBTORS ............................ 206
X I. M EDICAL D EBT .............................. 208
XII. CREDIT CARD DEBT .......................... 209
XIII. DEBTORS' PREVIOUS FILINGS ..................... 210
XIV. CHAPTER 13 PLANS .............................. 211
XV. FINANCIAL AND SOCIAL VARIABLES ARE NOT PRE-
D ICTIVE ... .. ... .. .... .. ... .. ... .... .... .. ... 2 13
XVI. DIVORCE AND BANKRUPTCY .................... 213
XVII. DEFICIENCIES OF THE ECONOMIC MODEL ......... 214
* Justice Joseph Weintraub Professor of Law, Rutgers University School of Law -
HeinOnline -- 43 Rutgers L. Rev. 185 1990-1991
R UTGERS LA W RE VIEW [Vol. 43:185
XVIII. PERSUADING DEBTORS TO SELECT CHAPTER 13 ... 221
XIX. RELUCTANT CREDITORS ........................ 223
XX. CONSUMER CREDITORS ........................ 224
A. Undersecured Creditors ................. 226
B. Conclusions About Consumer Creditors ... 229
C. Comparing the Consumer Creditors ...... 233
D. Creditor Legal Action ................... 235
E. Income to Debt Ratios; Mortgages ........ 235
F. Their Material on Consumer Credit Con-
cludes with Two More Fantastic State-
m en ts . ..... ... .... ... ... .... ...... .... . 236
G. Consumer Credit as Commercial Credit... 237
XXI. WHAT THEY CLAIMED TO HAVE LEARNED FROM
THEIR D ATA ................................. 238
XXII. OMISSIONS IN THE SWW BOOK ................ 241
For readers unfamiliar with bankruptcy jargon, the person fil-
ing in bankruptcy is called the debtor. Spouses may file together
in a joint petition. Filing in chapter 7 results in a liquidation of
the debtor's assets. The debtor surrenders his non-exempt assets
which are sold ana the proceeds paid over to his creditors. Ex-
empt property is retained by the debtor. What is exempt is deter-
mined by reference to varying state laws in about two-thirds of
the states, and by the debtor's selection of federal or state exemp-
tions in the rest. In 1981, when Sullivan, Warren and Westbrook's
sample of debtors filed for bankruptcy, about one-third of the
states had "opted out" by enacting legislation to foreclose use of
the comparatively generous federal exemptions.
Chapter 13 is the individual's reorganization. The debtor re-
tains all his property, exempt or not, but has to pay his creditors
at least some of his debts pursuant to the debtor's plan, which
must be approved and confirmed by the bankruptcy court.
I report my own interest as a part of this essay. I was deputy
director of the Commission on the Bankruptcy Laws, and am the
author of three of the empirical studies listed below.' I was also
1. The following abbreviations are used throughout this book review:
SWW: T.A. SULLIVAN, E. WARREN & J.L. WESTBROOK, As WE FORGIVE OUR DEBTORS:
BANKRUPTCY AND CONSUMER CREDIT IN AMERICA (1989).
NJ Debtors: Shuchman, New Jersey Debtors 1982-1983: An Empirical Study, 15 SETON
HALL L. R v. 541 (1985).
HeinOnline -- 43 Rutgers L. Rev. 186 1990-1991
1990] BOOK REVIEW
one of the reviewers of Sullivan, Warren and Westbrook's grant
proposal to the National Science Foundation and wrote in favor
of awarding their grant.
Most of their study replicates several earlier research publica-
tions. These are hardly mentioned. The writers make extravagant
and false claims to originality and priority of research. There ap-
pear to be serious errors in their use of statistical bases which
result in grossly mistaken functions and comparisons. Some of
their conclusions cannot be obtained even from their flawed find-
ings. The authors have made their raw data unavailable so that
its accuracy cannot be independently checked. In my opinion, the
authors have engaged in repeated instances of scientific
I. WHAT WAS DONE
This book is the report of an empirical study of personal bank-
ruptcy filings in 1981. The authors are two law teachers and a
demographer. They took information from 1529 files in Illinois,
Pennsylvania and Texas. The authors had this information coded
and tabulated, and used various statistical tests to analyze the
information taken from the files. They compared their results
with other data bases, largely federal figures of various kinds.
This book has been six years in the research and writing. The
authors' data were eight years old at publication. Nineteen
eighty-one was a recession year. Unemployment has decreased
and remained low in the eight years since then. Interest rates
were very high in 1981, but were sharply lower within two years
and have remained quite low relative to 1981. There were sub-
stantial amendments to the Bankruptcy Code in 1984. These fac-
tors make it almost impossible to gauge the present value of the
authors' data. Only a study over time and an impact analysis of
the 1984 amendments could show whether their findings, if accu-
Nine Sts.: Shuchman, The Average Bankrupt: A Description and Analysis of 753 Per-
sonal Bankruptcy Filings in Nine States, 88 CoM. L.J. 288 (June/July 1983), errata at 88
COM. L.J. 395 (Aug./Sept. 1983).
GAO: U.S. GENERAL ACCOUNTING OFFICE, BANKRUPTCY REFORM ACT OF 1978-A BEFORE
AND AFTER LOOK (1983).
CRC: CREDIT RESEARCH CENTER, CONSUMER BANKRUPTCY STUDY (1982).
CT: Shuchman & Rhorer, Personal Bankruptcy Data for Opt-Out Hearings and Other
Purposes. 56 AM. BANKR. L.J. 1 (1982).
Brookings: D. STANLEY & M. GIRTH, BANKRUPTCY: PROBLEM, PROCESS, REFORM (1971).
HeinOnline -- 43 Rutgers L. Rev. 187 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
rate for 1981, remain descriptive now.
II. KNOWLEDGE AND LEGISLATION
Throughout their book, the authors raise the issue of the
knowledge which the Congress did not have when the Bankruptcy
Code and its amendments were enacted. The authors' pleas for
more empirical information are well taken in principle but are,
variously, unrealistic, grossly exaggerated, and based on excited
but inaccurate apprehensions. I too have contended for more em-
pirical information about the workings of the legal process and its
effects. But this problem is not helped by their accusations, for
example, that the Bankruptcy Code was legislated in such igno-
rance as to call it "a vacuum of fact." 2 They sound like converts
with all the breathless enthusiasm of a new revelation. Finally,
the authors tell us:
Bankruptcy reform requires hard information about the players.
Without it, policymaking is little more than reading tea leaves,
responding to political pressures, and acting out unexamined
The authors intend to provide that required "hard informa-
tion" about debtors and their creditors to which they refer, and
they claim to have done so in this book for the first time. They
fail because the claimed revelations which are presented have
been widely known for several years and were well known to the
Congress during the latest round of amendments to the Bank-
ruptcy Code in 1984.
III. SOME OF THE AUTHORS' CLAIMS
The authors began this work, they tell us, "because of our reali-
zation that so little was known about this important legal and
social phenomenon."'4 In particular, they point out that
"[blecause existing data were inadequate, we had to collect data
on the characteristics of bankrupt debtors, their assets, their lia-
bilities, their jobs, their marital status, whether they were home-
owners, and so forth. ' By the middle of 1983, when they began,
2. SWW, supra note 1, at 7.
3. Id. at 280.
4. Id. at 338.
5. Id. at 342.
HeinOnline -- 43 Rutgers L. Rev. 188 1990-1991
1990] BOOK REVIEW
all the other studies listed above,6 except for NJ Debtors, had
been published. To claim that they began their work because so
little was known and what was known was inadequate is not only
untrue, but is also something of a put-down to other toilers in
this field. The authors make up for that, however, by virtually
replicating the previous studies and confirming most of their
findings, and doing that now with data from nine years ago.
The little new information in this book is of minor conse-
quence.' The most interesting different findings are that 52% of
the authors' sample of all debtors in bankruptcy are homeowners,
and that chapter 7 debtors own homes about as often as do chap-
ter 13 debtors." These findings are in sharp disagreement with
those of earlier studies with data from about the same time,9
which show chapter 7 debtors' home ownership at 16% to 28%,
and with the GAO study which concluded that chapter 13 debtors
owned homes twice as often as those in chapter 7, about 50% to
25%. Had the authors compared their findings with others' of
about the same time, they might have paid attention to why these
other studies apparently all went astray and in the same
The authors' Introduction states what they believe they have
done: "We . . .explore the impact of judges and lawyers on the
practical implementation of bankruptcy."" This claim is neither
much nor well supported in the book. The authors relate that
they "interviewed several different people with different connec-
tions to the bankruptcy system, including bankruptcy judges,
trustees [and] lawyers,"' 2 but state that "our use of the interview
data is limited." 8 There was no systematic questioning and the
authors present no data based on these interviews.
6. See supra note 1.
7. The authors make much of their findings that chapters 7 and 13 debtors appear to
have about the same income to debt ratios irrespective of different exemption laws, con-
tending that this is a direct test which disproves what they term the economic model.
They do not mention that GAO had previously found that "chapters 7 and 13 debtors
have similar income to debt relationships." GAO, supra note 1, at 53.
8. SWW, supra note 1, at 129.
9. See infra Table 1.
10. "In studying a problem, every scientist must assimilate and assess the results of
previous studies of the same problem." FAmNzss iNEMPLOYMENT TESTING 120 (J. Hartigan
& A. Wigdor 1989).
11. Sww, supra note 1, at 10.
12. Id. at 351.
13. Id. at 352.
HeinOnline -- 43 Rutgers L. Rev. 189 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
They claim also to "examine the increasing fragility of the con-
sumer credit system,""' but the authors do not tell the reader
what they mean by "fragility" in this context. They make no ef-
fort at showing whether the consumer credit system was more or
less fragile earlier or later than their references. Without such in-
formation, the question whether its fragility is increasing remains
unanswered. There is no fragility from the lenders' standpoint.
Consumer credit remains for most large lenders a very profitable
part of their portfolios.
The authors claim that they will test "the unspoken premises
of the bankruptcy debates against hard data about how bank-
ruptcy actually works."' 5 I know of no unspoken premises and the
authors tell us of none.
They say the reader "will learn about the pathology" of the
consumer credit system. 6 From this book a reader could not find
out what distinguishes a healthy from a pathological consumer
credit system, assuming those adjectives have specifiable
The scientific flavor of the book is neatly illustrated at the very
beginning. "This book is based on . . .the largest study of con-
sumer bankruptcy ever undertaken. It required . . .a sample of
2400 bankruptcy pbtitioners . . . .,'v That claim is reducible to
the examination of 1529 bankruptcy files. The 2400 petitioners is
reached by counting as two joint filings by spouses, which all
other researchers count as one. Both the CRC and the GAO stud-
ies are based on larger samples. Including those debtors not inter-
viewed (SWW also did not interview debtors), the CRC sample is
2249 files. The GAO sample is 1699 files, including 415 mailed
responses to questionnaires.
IV. PROBLEMS IN THEIR APPROACH TO EMPIRICAL RESEARCH
There is an anti-scientific bias to the authors' work. The four
other nearly contemporaneous empirical studies of bankruptcy
use some different methods," are from some other areas and
14. Id. at 8.
15. Id. at 7.
16. Id. at 3.
17. Id. at 4.
18. "[M]ultimethod strategy allows weaknesses in one design to be corrected by
strengths in another. When outcomes from different methods converge, greater confidence
in the research findings is the result." Diamond, Methods for the Empirical Study of Law,
HeinOnline -- 43 Rutgers L. Rev. 190 1990-1991
1990] BOOK REVIEW
cover time periods other than 1981 filings," but are temporally
close to the present sample. The findings of these studies are not
mentioned and are not compared.""
We deal here with the pervasive and vital problem of the exter-
nal validity-the generalizability-of the SWW study's findings
and conclusions."' External validity is often established by repli-
cation. Using these other studies provides several advantages: we
see these matters over more time and in other places and we can
compare the same or similar research done by other investigators.
Also, there is the possibly great advantage resulting from the in-
terviews of debtors conducted by CRC and NJ Debtors. The in-
formation from those interviews which supplemented their archi-
val studies, even if less reliable than the bankruptcy files, could
have been of great importance to the SWW study.
The only comparisons the authors undertake are with those of
the Brookings Institution study of some twenty-five years before
which was under a very different bankruptcy law and much
changed debtor-creditor laws in most states. Those comparisons
in LAW AND THE SOCIAL SCIBNCBs 637, 643 (1986) (citation omitted).
Researchers have lately been applying the statistical techniques termed "meta-analysis."
Meta-analysis "is like an ordinary scientific review of research, except that ordinary re-
views provide a qualitative-and often subjective-assessment of a few studies; meta-anal-
ysis. .. promises a quantitative synthesis of all available data." Mann, Meta-Analysis in
the Breech, 249 Sc. 476-80 (1990).
Other researchers confirm that meta-analysis "produces more robust evidence than any
single study. The convergence of evidence produced under differing conditions helps to
ensure that the effects observed are not the inadvertent result of some [other unrecognized
factors]." FAmNuss IN EMPLOYMENT TESTING, supra note 10, at 120.
19. Two of the important characteristics of legal behavior are variability (1) across set-
tings and (2) over time. Diamond, supra note 18, at 660. The authors' one slice in time
approach cannot take account of the second threat to generalizability. Part V,infra, shows
that their site selection probably does not protect against the first.
20. The exception is their attack on the CRC estimates of the feasibility of compulsory
chapter 13 paybacks.
21. See generally T. CooK & D. CAMPBELL, QUASI-ExPERMENTATION 70-80 (1979).
22. A partial list of the more important changes since the Brookings research follows:
Since 1970, the bankruptcy courts, not the state courts, have had exclusive jurisdiction to
decide the dischargeability of debts. This jurisdictional change had the desired result of
sharply reducing the frequency of such actions and reaffirmations as well. The 1978 Bank-
ruptcy Code made it even more difficult and slightly hazardous for the creditors to obtain
reaffirmations. These are both apt to result in considerable empirical changes. Despite
these two major changes in the past 20 years, the authors compare their apparently frag-
mentary data on reaffirmations with Brookings "25 years earlier." SWW, supra note 1, at
The 1978 Code in its important aspects created a federal exemption and resulted in
larger state property exemptions in most states.
HeinOnline -- 43 Rutgers L. Rev. 191 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
are of little value for most purposes.
The authors know the importance of confirming the results of
other studies and the value of replication. 3 But the text of the
SWW study purports neither to confirm nor to rebut what other
studies have found. That is because the authors do not discover
new facts. They rediscover facts previously known. According to
Maier's Law, "the important thing when rediscovering facts is not
to mention their former existence. Reincarnations must always be
in a form different from the former state and, of course, should
have a different name."" The most efficient method for disposing
of others' facts and findings is by failing to report them. This
method, if skillfully handled, permits a selection process so that
only facts and findings that are material and important are re-
ported. Naturally it takes critical individuals such as the authors
to determine what is worth reporting and what need not be
The authors' selection of Illinois, Pennsylvania and Texas was
"because they varied in interesting ways that we believed might
be important to bankruptcy outcomes. 2 6 Without explanation
they tell us that these states have diverse economies. Illinois and
Pennsylvania had higher unemployment rates in 1981 (8.5% and
8.4%) than did Texas (5.3%). The crucial variable for the au-
thors' study is the wide variation in these states in the amount of
property which debtors can exempt from creditors.
It is unclear to me how these three states were selected. We are
Chapter 13 was drastically changed to encourage its use by individual debtors and indi-
vidual owners of small business firms.
A number of landmark Supreme Court decisions, starting with Sniadach v. Family Fin.
Corp., 395 U.S. 337 (1969), dramatically changed the state laws of prejudgment property
attachment and wage garnishment.
The Consumer Credit Protection Act, 7 UNIFORM LAws ANNOTATED (Master ed.
1985)(1968 Act), 15 U.S.C. § 1601 et seq. (1988), adopted in 1968, and the Uniform Con-
sumer Credit Code, as adopted, greatly changed the laws on exemptions from wage gar-
nishment and deficiency judgments following personal property repossession.
23. See Sullivan, Warren & Westbrook, Limiting Access to Bankruptcy Discharge:An
Analysis of the Creditors' Data, 1983 Wis. L. REv. 1091, 1138.
24. "Many superfluous data are collected because it is too difficult or confusing or un-
convincing or unglamorous to assemble and examine what is known already." FAIRNESS IN
EMPLOYMENT TESTING, supra note 10, at 120.
25. See Maier, Maier's Law, 15 AM. PSYCHOLOGIST 208, 211 (1960).
26. SWW, supra note 1, at 18.
HeinOnline -- 43 Rutgers L. Rev. 192 1990-1991
1990] BOOK REVIEW
not told what criteria were used to reject other possible candidate
states. The authors should have tried to maximize the heteroge-
neity of the states with respect to what might be the important
A significant factor, which SWW did not investigate, is the se-
verity of wage garnishment laws in the states. Brookings had
found that "bankruptcy rates in various states appeared to be re-
lated to their garnishment laws.""7 That increased wage exemp-
tions from garnishment result in reduced bankruptcy rates was
confirmed in a Bankruptcy Commission study that examined the
impact of the federal law requiring of the states a minimum ex-
emption from wage garnishment. 2 This relationship was tested
again for the years just before and after the Bankruptcy Reform
Act when it was again confirmed that bankruptcy rates were con-
sistently lower in states that prohibited or severely restricted
wage garnishment. The relationship was more pronounced in 1980
than in 1978.2s
To the extent wage garnishment is a contributing factor in the
decision whether to file, these three states are not good choices.
Pennsylvania and Texas prohibit wage garnishment altogether, 0
while Illinois had a one-third more generous exemption for wages
than did federal law. Without a better selection of states it is also
difficult to detei'nine whether wage garnishment law bears any
relation to the choice of chapter 7 or 13. Another resulting prob-
lem for the authors' research is that these states are not a good
test of what they call the economists' model. '
These questions about site selection go to the most basic issue:
Are their findings generalizable? Would they hold true in other
states? And at other times? One way to assay this is to compare
their findings with the results of research conducted by others. It
is common in such research to draw comparisons across indepen-
dent studies previously done. This kind of replication, if the re-
sults are consistent, has a cumulative effect. Confidence is en-
27. Brookings, supra note 1, at 28-32, 49.
28. Shuchman & Jantscher, Effects of the Federal Minimum Exemption from Wage
Garnishment on Nonbusiness Bankruptcy Rates, 77 CoM. L.J. 360 (Nov. 1972).
29. Peterson & Aoki, Bankruptcy Filings Before and After Implementation of the
Bankruptcy Reform Law, 36 J. ECON. & Bus. 95, 103 (1984). See also id. at 101 n.4 for
other similar studies.
30. Of the 50 states, only these two states and Florida had a complete exemption of
wages from garnishment.
31. See infra Part XVII.
HeinOnline -- 43 Rutgers L. Rev. 193 1990-1991
194 RUTGERS LAW REVIEW [Vol. 43:185
hanced in one's findings, especially if the other studies are
HeinOnline -- 43 Rutgers L. Rev. 194 1990-1991
1990] BOOK REVIEW
m 0 Lo
C*4 cq -4
G& G& f* G&
eq) () eo ~- r-
04 ~ ~ 0
HeinOnline -- 43 Rutgers L. Rev. 195 1990-1991
196 RUTGERS LAW REVIEW [Vol. 43:185
1. Mean income reported in table G of study adjusted to include 21 cases with zero income. Nine
Sts., supra note 1, at 10.
2. CRC, supra note 1, at 78.
3. GAO, supra note 1, at 23.
4. Debtors owned a home in 167 of the 753 caes. Nine Sti., supra note 1, at 47.
5. Debtors owned a home in 332 of 1199 cas. CRC, supra note 1, at 78.
6. "A greater percentage of chapter 13 debtors were homeowners compared to chapter 7 debt-
ore--50 percent compared to 25 percent respectively." GAO, supra note 1, at 49. GAO's responses to
mailings revealed more homeownership, but the ratio of chapter 13 homeowners to chapter 7 homeown-
ers remained about two to one (chapter 13 - 66%; chapter 7 - 32%; N - 415). Id. at 20-21.
7. Income in the calendar year before the petition was filed was not reported for 78 cases and was
zero for 21 cases. These cases were classified as unemployed for purposes of this comparison.
8. The percentage of sample that reported positive wage income on the petition for year before
filing. Only 78.6% of sample had at least one person employed at the time the petition was filed. CRC,
aupra note 1, at 33.
9. GAO, supra note 1, at 34.
10. CRC, supra note 1, at 78.
11. GAO, supra note 1, at 48.
12. The average total debt was calculated by summing the totals of mortgage, unsecured, nonmort-
gage secured, and priority claims and dividing by 753 cases.
13. CRC, supra note 1, at 78.
14. GAO, supra note 1, at 51.
15. Total debt minus average mortgage debt (for whole sample).
16. Average total debt minus an average debt secured by home/mobile home of $9561. GAO, supra
note 1, at 27.
17. This figure differs from the 132% figure reported in Nine Sts., supra note 1, at 56, because of the
recalculation of average nonmortgage debt and the use of average income from previous year rather
than the average income for the previous two years. Both of these changes necessary to make a valid
comparison with the other two studies.
18. For 350 cases filed using the federal bankruptcy exemption law only.
19. Weighted average for 1980 and 1981. N.J. Debtors, supra note 1, at 658 table 19.
20. Id. at 575.
21. Id. at 560 table 21. Total N,186 of which 173 were employed.
22. $4129 personal property + 58677 real property (averages for those reporting equity). Id. at 583.
23. Id. at 565.
24. SWW, supra note 1, at 64 table 4.1.
25. Id. at 129.
26. Id. at 86.
27. Id. at 64 table 4.1.
29. Id. at 69.
30. Id. at 64 table 4.1.
31. Id. at 64 table 4.1, 66-69.
32. Id. at 260 table 13.7.
HeinOnline -- 43 Rutgers L. Rev. 196 1990-1991
1990] BOOK REVIEW
Table 1, taken from the CRC publication, "The Monitor," and
expanded, is the beginning of such an effort. Perhaps, by careful
comparison with the four other contemporaneous empirical stud-
ies of personal bankruptcy, I can help further the research effort
by showing that the results of the SWW replication are consis-
tent, and thereby add strength to their findings. Where their rep-
lication does not lead to similar results, perhaps the differences
can be attributed to varying research techniques for which proba-
ble explanations can be tested by others. Where their replication
failed, those differences (as in home ownership) may suggest new
questions for investigation.
VI. CRITICISMS OF THE ADMINISTRATIVE OFFICE STATISTICS
The authors' very strongly stated criticisms of the Administra-
tive Office ("A.O.") statistics are (1) its one-time decision to
count joint petitions as two filings; and (2) that the A.O. cannot
accurately decide which filings are business-related and which are
The one-time blip was obvious and persons involved with these
figures knew about the counting of joint filings as two. The 1978
Bankruptcy Code was a new law on joint filings which allowed
spouses to have a single case with one filing fee. Before 1979, the
estates of spouses were treated as separate cases with two filing
fees paid; they were usually then consolidated for processing. In
the event, both sets of figures were published by October, 1981.31
Also, concerned persons have known for years that 10% or
more of the personal bankruptcy filings were business-related.
There was abundant evidence in the journal literature on the pro-
portion of business-related personal bankruptcy filings.3 4 The au-
thors say that "[a]ny analyses that depend on that [personal-bus-
32. SWW, supra note 1, at 40-41. The authors claim that they discovered how mislead-
ing the A.O.'s gross statistics are as their research developed. Id. at 338. That must be a
mistake. They started their research in mid-1983. Well before that time the matter of joint
filings as two estates had been clarified, with both sets of figures published. There had
also been discussion of the business-related personal bankruptcies in the literature from
33. See Bankruptcy Reform Act of 1978 (Future Earnings) Hearings Before the Sub-
comm. on Courts of the Senate Comm. on the Judiciary, 97th Cong., 1st Ses. 187, 190
table III (1981) [hereinafter Future Earnings Hearings).
34. See infra table 3. A separate category would be helpful to researchers and probably
to the Congress, as was suggested in Nine Sts., supra note 1, at 306.
HeinOnline -- 43 Rutgers L. Rev. 197 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
iness] classification are fatally flawed." 5 This is an exaggerated
claim, bold in appearance but meaning very little. There are not
apt to be recent analyses that depend on that classification with-
out making some allowance for or relying on a sampling of per-
VII. PROBLEMS IN THEIR METHODS
For the most part, the authors' sample (all from 1981) of 938
chapter 7's and 591 chapter 13's are treated as one group. This is
a departure from previous studies of this type, and may distort
some of their findings. For one small example, the average legal
fee in the SWW sample is five hundred dollars37 which, because it
includes both chapter 7's and 13's, is much higher than CRC's
four hundred" and NJ Debtors four hundred forty dollars.
There are, however, two more serious problems with their
A. 1981 Incomes of Chapter 7 Debtors
It is not clear to me how the authors arrived at the income
figures for their chapter 7 debtors in 1981, the year of their sam-
ple filings. For their sample of chapter 7 debtors, they would ob-
tain from the 1981 files incomes for 1980 and 1979.0 For their
chapter 13 debtors they would obtain present-not an-
nual-income, defined as "gross wages, salary or commissions per
pay period" for the debtor and spouse, even if not a joint petition,
and "gross income for the last calendar year." '
In order to arrive at the annual incomes for 1289 debtors, 2 the
35. SWW, supra note 1, at 330.
36. If an analysis merely took account of the observation in Brookings, supra note 1, at
47, that 13% of its sample of personal bankrupts had been in business, it would have been
in line with the later empirical studies of the 1980's, although far from the authors' anom-
alous finding of 20%. See infra Part IX, table 3.
37. SWW, supra note 1, at 43 n.19.
38. CRC, supra note 1, at 52.
39. NJ Debtors, supra note 1, at 587. Later in the book the authors break down the
average fees. Chapter 13's are $535 and chapter 7s are $459. SWW, supra note 1, at 250.
40. United States Bankruptcy Court Official Form No. 7 (for chapter 7), questions 2.d
and 2.e, ask about income "during each of the 2 calendar years immediately preceding the
filing .... "
41. United States Bankruptcy Court Official Form No. 10, question 1.f(11).
42. See SWW, supra note 1, at 64 table 4.1.
HeinOnline -- 43 Rutgers L. Rev. 198 1990-1991
1990] BOOK REVIEW
authors would have to have used at least 698 chapter 7 debtors"
whose incomes would have been given for 1980 but not for the
year of filing, 1981. Thus, they would also have to have used 1980
incomes or estimated 1981 incomes for chapter 13's and 1980 in-
comes for chapter 7's. Yet throughout the authors use 1981
benchmarks." Readers are apt to be misled by this device and the
resulting mistaken comparisons to federal benchmarks,'46 and
debt to income ratios. 4
B. Their Tenuous Inference
The lower incomes of debtors help explain their financial
stress but another piece of information puts the income data
into context. The family income of bankrupt debtors is also
spread among more people. In the United States in 1981, mean
household size was 2.7. Mean family size among debtors in our
sample was about 3.4, or nearly one person more. 7
The problem here and throughout most such data in the book
is that, although not stated in the text, family size is taken only
from the files of chapter 13 debtors, not from chapter 7 debtors.
The authors explain in a footnote: "Only debtors in Chapter 13
are required to list all economic dependents, so our report on
family size for all districts is based on Chapter 13 debtors."" Lo-
cal rules in the Western District of Texas, however, required all
debtors, chapter 7's and 13's alike, to give family size. "In that
district the mean family size for Chapter 7 debtors was not signif-
icantly different from the family size for the Chapter 13 debtors
... .This suggests that the computed family size for Chapter 13
43. Their total sample is 938 chapter 7's and 591 chapter 13's. From the 1289 debtors
used for "Family Income" in SWW, supra note 1, at 64 table 4.1, deducting the 591 chap-
ter 13's leaves 698 chapter 7 debtors.
44. See, for example, the references to income data. SWW, supra note 1, at 65.
Since this section was written, Dr. Sullivan has advised me by letter that they "used the
1980 income figures from Chapter 7 as a proxy for current [i.e., 1981] income." Letter
from Teresa A. Sullivan to Philip Shuchman (Feb. 6,1990) [hereinafter Letter] (discussing
statistics in SWW)(emphasis added).
45. See infra Part VIII.
46. See infra Part XVII.
47. SWW, supra note 1, at 65. GAO's chapter 13 debtors seem to have significantly
larger households than their chapter 7's. GAO, supra note 1, at 14. NJ Debtors' interviews
of 55 chapter 7 debtors found an average family size of 3.36. NJ Debtors, supra note 1, at
552 table 12.
48. SWW, supra note 1, at 81 n.13.
HeinOnline -- 43 Rutgers L. Rev. 199 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
debtors is not different from the family size for Chapter 7
Using chapter 13's to tell us about family size and family in-
comes, and the family income to debt ratios of chapter 7's, is a
hazardous undertaking.50 Relegating that information to a foot-
note is something else. It takes a careful reader of footnotes to
discover that inference. Readers might not attempt to assay
whether particular similarities in seventy-three chapter 13's and
seventy-five chapter 7's in the Western District of Texas would
extend to the rest of SWW's sampled districts and would be
generalizable to the whole nation's debtors.
This shaky foundation also supports other related inferences.
The authors advise that "[blecause of the important financial in-
formation they reveal, we examine the debt/income ratios for
each of the debtors in our sample." 5 This is supported elsewhere
in a footnote. "For members of families, income is added for the
whole family, based on the information provided in the bank-
ruptcy files. Debt is also compiled by family, so that the debt/
income ratio describes the economic position of a whole family
whenever that information is available."5 " Family size, however,
was not given in chapter 7 petitions. It was available only for
chapter 13's and, irk some form, for those seventy-five chapter 7's
in the Western District of Texas. Family members who were not
dependents or contributors were not included in the schedules
filed in either chapter.
The authors inform the reader that in joint petitions the
spouses must list all income from all sources. This does not nec-
essarily mean family income since the authors do not know
whether other family members existed and whether they earned
any income in the two years preceding the year of filing in bank-
ruptcy. For example, table 4.154 does not describe family income
unless all other members of the family received income and paid
50. The authors might have tried small similar samples after the Bankruptcy Amend-
ments and Federal Judgeship Act of 1984, 28 U.S.C. § 151 et seq. (1988), which then used
Official Form 6A (based on the chapter 13 form) to get some broader and better notion of
how chapter 13's and 7's compared in these regards. They did not attempt any tentative
51. SWW, supra note 1, at 73.
52. Id. at 83 n.27.
53. Id. at 78 n.4.
54. Id. at 64.
HeinOnline -- 43 Rutgers L. Rev. 200 1990-1991
1990] BOOK REVIEW
it over to the debtors. This table should not include chapter 7
d'&btors, even joint petitioners, because there were no questions
about the family or its size in 1981. Only if a member of the fam-
ily paid the filers something in the preceding two years were they
required to respond. None of the chapter 7 debtors was required
to say anything of any income for the year of filing. If the infor-
mation on family income in table 4.1, with a sample size of 1289,
came from all 591 of their chapter 13's, the authors still would
have been guessing about the family size of the 698 chapter 7's
included, or more than half of this subset.
VIII. INCOME CHANGES
The authors "explore the possibility that debtors may have suf-
fered income interruption or sharp earning declines . . . ."" In-
formation on this very subject had just been provided in Nine
Sts." The authors think that "[i]ncome volatility is an important
factor in consumer bankruptcy. 5 6 7 Their rediscovery of changes in
income comes reincarnated with a different name, "income vola-
tility." When a person becomes unemployed or is working fewer
hours, his or hen income declines. If that is unexpected or unan-
ticipated, debts previously incurred cannot be repaid. Some of
those persons who cannot pay will file in bankruptcy. (Some
think newly re-employed persons are apt to file.)
One of the two reasons the writers give for the great increase in
personal bankruptcies is "income volatility."" I have taken their
figures on income changes in the two calendar years preceding
bankruptcy and prepared Table 2 below which compares their
numbers with those of Nine Sts. and NJ Debtors, and with some
percentages from CT.
55. Id. at 98.
56. Supra note 1. It was provided earlier in CT, supra note 1, at 11. "Half of [the] files
showing income for both preceding calendar years reported an increase in income....
Twenty percent of the group . . . showed declining income, and the remainder stayed
about the same."
57. SWW, supra note 1, at 333.
58. Id. at 12. Later they hedge. "Unfortunately, we have no clearly comparable national
statistics on income volatility or its trends. But our data certainly suggest that the increase
in bankruptcies is linked to an increase in income interruptions among debtors." Id. at
HeinOnline -- 43 Rutgers L. Rev. 201 1990-1991
202 RUTGERS LAW REVIEW [Vol. 43:185
% of Cases Av $ Increase % Increase
SWW c.50% $3500 +70%
Nine Sts. 57% $3289 +27%
NJ Debtors 56% $3831 +25%
% of Cases Av $ Decrease % Decrease
SWW 25% $5000+ -37%
Nine Sts. 23% $3663 -25%
NJ Debtors 26% $3675 -21%
% of Cases No Change No Change
SWW c.20% + - -
Nine Sts. 15% - -
NJ Debtors 18% - -
CT c.30% - -
SWW, supra note 1, at 98-99.
Nine Sts., supra note 1, at 293 tables H-K.
NJ Debtors, supra note 1, at 560 table 21.
CT, supra note 1, at 11.
The authors are surprised that more debtors had incomes
which had gone up and that the gains were more than the losses.5 9
This would not have been surprising to them had they read Nine
Sts. or NJ Debtors. Although the authors' findings are of greater
magnitude, the directions and proportions are similar.
There was some interview evidence, in person and by mail,
which was generally confirmed, although not mentioned by SWW.
The CRC respondents were direct and explicit in this matter.
Forty-four percent stated that an important reason for filing in
chapter 7 bankruptcy was being "laid off or fired from work"
(27%), and a "cutback in hours worked" (17%)."0
In ranking the factors that led to filing bankruptcy, GAO found
that "periods of unemployment contributed to a great extent ac-
cording to 36% of chapter 7 debtors and 34% of chapter 13 debt-
59. "There were more debtors whose incomes had risen than those who had lost ground,
and the gains were even bigger than the losses. This was an utterly counterintuitive find-
ing." Id. at 99.
60. I CRC, supra note 1, at 38 exhibit 3-3.
HeinOnline -- 43 Rutgers L. Rev. 202 1990-1991
1990] BOOK REVIEW
ors . . .
Nine Sts., which had 42% of its sample of filings from 1980 and
the last four months of 1979, reported on incomes for calendar
year 1978. The frequencies of changes are comparable to those
which the SWW study finds unusual. NJ Debtors, which includes
four years of incomes from 1978 through 1981, also shows closely
comparable frequencies of change. That two additional years of
debtors' incomes in different places show similar frequencies of
changes suggests that the SWW findings are not unusual.
The authors say that "[n]early 62% of the debtors had a
change in income, up or down, of more than 10% from 1979 to
1980. Even for turbulent times, these figures portray highly vola-
tile income streams, making a mismatch between debts and in-
come likely."6 NJ Debtors had shown average income increases
for the entire sample of debtors for the years 1978 through 1981:
1981 = 10.4%"
In three of the four years, average total incomes increased by
more than 10%. But the average annual increase in the cost of
living as measured by the Consumer Price Index was more than
10% during those four years. Hence, income increases of at least
10% hardly matched the increases in the cost of living. Many la-
bor contracts use the Consumer Price Index as the basis for wage
changes. It does not appear that the frequency or magnitude of
the changes in income that the authors report were unusual or
should be characterized as "highly volatile income streams." Also,
since all the empirical studies agree that about 70% or more of all
debtors' incomes stay about the same or increase in the two years
before bankruptcy filing," ' there is no basis for their conclusion
that the changes in income make "a mismatch between debts and
income likely"" which might lead to bankruptcy. If anything, the
authors should have drawn the contrary inference.
The empirical studies compare debtors' incomes to various
benchmarks such as the Bureau of Labor Statistics ("BLS")
61. GAO, supra note 1, at 15.
62. SWW, supra note 1, at 98.
63. NJ Debtors, supra note 1, at 560.
64. See supra Table 2.
65. SWW, supra note 1, at 98.
HeinOnline -- 43 Rutgers L. Rev. 203 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
budgets. For their sample of chapter 7 debtors the SWW study
uses 1981 benchmarks, although they have only 1980 incomes for
those debtors. This is apt to distort their comparisons because of
the increases in the debtors' incomes from 1980 to 1981. The NJ
Debtors figures"6 may illustrate the probable extent of SWW's
Average Income of Debtors $16,251 $19,010
BLS Lower Living Budget $14,898 $15,705
Had the writers used both the income and BLS benchmark fig-
ure for 1981, the ratio of these figures would be .8261 ($15,705
divided by $19,010). The ratio arrived at by SWW was .9664
($15,705 divided by $16,251). The difference, i.e., the extent of
SWW's error, is .1403 (.9664 - .8261 = .1403), which translates
into a result for the SWW analysis which is about 17% lower than
the correct ratio of income to the BLS benchmark. The authors'
error may be much larger because half of their sample showed
average increases of 70% while only a quarter showed decreases
averaging about 37%.
If, by "income volatility," the authors meant unemployment
variables (rate of unemployment and average number of hours
worked per week), those data are available regionally and nation-
ally for almost the whole post-war period, from 1945. They do not
even examine those figures over time, nor do they attempt to find
out what relationship exists, if any, between employment condi-
tions and bankruptcy.'
About all that the authors can say with confidence is that their
data confirm what most previously published studies found. Most
bankrupt debtors are employed, but their unemployment rate
(10% to 20%) is higher than the whole population of wage earn-
ers. Since they provide no adequate measure and no history of
"volatility," they cannot say that an increase in personal bank-
66. NJ Debtors, supra note 1, at 558 table 19.
67. Peterson and Aoki found that much of the increase in the rate of bankruptcies from
1978 to 1980 could be accounted for by changes in the level of unemployment or the aver-
age work week. Unemployment rose and weekly hours of work declined during that time.
The effects of these factors seemed to them to be greater in 1980 than in 1978. Peterson &
Aoki, supra note 29, at 95.
Other research suggests that this relationship may be more complex than simple models
can describe. See Luckett, Personal Bankruptcies, 74 Fed. Res. Bull. 591, 595 table 1
HeinOnline -- 43 Rutgers L. Rev. 204 1990-1991
1990] BOOK REVIEW
ruptcies is due to "increased volatility."
IX. BUSINESS-RELATED PERSONAL BANKRUPTCIES
The title and the content of this material make clear the work-
ings of Maier's Law. What had previously been called business-
related personal bankruptcies, taken to be the result of small bus-
iness failures, are now called "entrepreneurs," those who for-
merly, or at filing, were self-employed in some business venture.
Table 3 below puts the authors' figure in perspective.
Frequency of Business-Related Personal Bankruptcies
SWW CT Nine Sts. CRCU NJ Debtors
20%1 13%2 12%3 15% 4 11% 5
1. SWW, supra note 1, at 111.
2. CT, supra note 1, at 19.
3. Nine Ste., supra note 1, at 289.
4. CRC, supra note 1, at 38 exhibit 3-3.
5. NJ Debtors, supra note 1, at 553 table 13.
The SWW study has a larger group of what were known as
"former business debtors"' or "former DBA's."' 0 A likely reason
for their higher frequency is that the authors' definition of this
group apparently included persons who said they were self-em-
ployed at filing or during either of the two years before filing,
regardless of the debts scheduled.' In CT, Nine Sts. and NJ
Debtors, the decisive factor was whether the debts appeared to be
business-related, in which case the business failure probably ac-
counted for the bankruptcy filing.7 Definitions are neither good
68. Fifteen percent of CRC's respondents reported that their reason for chapter 7 bank-
ruptcy filing was that their "own business failed." I CRC, supra note 1, at 38 exhibit 3-3.
69. Brookings, supra note 1, at 44.
70. CT, supra note 1, at 19.
71. SWW, supra note 1, at 111.
72. "The filings are labeled personal ... because the proprietors were personally liable
for debts of the business and most of these business failures were of self-employed persons
and couples engaged in a variety of very small commercial enterprises and occupations.
The business failures in the form of personal filings are ordinarily evident from the sched-
ules listing the claimants (in particular, amounts of debt and types of creditors) and the
stated employment of the debtors." Nine Ste., supra note 1, at 288.
Business-related filings were identified "from the size of the debts and the identity of
the creditors, as well as from the debtors' statements of occupation." NJ Debtors, supra
note 1, at 588.
This is one of the authors' findings which is not compared even to Brookings. Theirs is
HeinOnline -- 43 Rutgers L. Rev. 205 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
nor bad, but the SWW definition is not useful because it pre-
cludes comparison with other empirical studies. Again, the gener-
alizability of this finding might have been gauged by the other
studies, which were conducted by different researchers, in other
places, and at the same and different times. There is another
problem in comparability created by the SWW study, which com-
bined chapter 7's and chapter 13's, but did not separately calcu-
late the chapter 7's for this purpose. All the other studies ex-
amined only samples of chapter 7 debtors, 3 except for GAO, which
had samples of both calculated separately.
The self-employed earned about the same amount (average
$17,700) as did the ordinary wage earners (average = $15,600),
with the difference reported by the writers as not statistically sig-
nificant. 4 Nine Sts. also showed about the same incomes for per-
sonal bankruptcies and business-related bankruptcies (all in
chapter 7) with but one exception. 7
X. FEMALE DEBT6RS
The authors' narrative on "Women and Bankruptcy" gives us
no new information of any significance. The SWW sample has
fewer females filing alone. Table 4 below compares the SWW
count of females filing alone with that of Nine Sts. and NJ
Females Filing Alone
SWW Nine Sts. NJ Debtors
% of all filings 17% 31% 24%
% of single filers 40% 47% 44%
SWW, supra note 1, at 149.
Nine Ste., supra note 1, at 289.
NJ Debtors, supra note 1, at 553 table 13.
not the definition used by Brookings: "A bankrupt who is not engaged in a business ven.
ture at the time the bankruptcy petition is filed but whose financial problems result from
earlier involvement in business." Brookings, supra note 1, at inside cover. Brookings found
these "former business" debtors were about 13% of their liquidation (now chapter 7)
cases. Brookings, supra note 1, at 44.
73. See GAO, supra note I.
74. SWW, supra note 1, at 114.
75. Nine Sts., supra note 1, at 293 tables J, K.
HeinOnline -- 43 Rutgers L. Rev. 206 1990-1991
1990] BOOK REVIEW
This discussion of women's occupations relates the generally
accepted findings that women are paid less than men and that
the differences in occupation do not account for the differences in
male and female incomes." These differences in earnings, they
find, are reflected in their sample of bankrupt debtors. NJ Debt-
ors shows the archival information from the files on occupations
and incomes for females." The four largest categories of female
occupations and incomes (for "last year") are compared with that
of males in Table 5 below.
Occupational Category Males Females
Clerical $ 9,699 $12,028
Service $14,725 $10,269
Professional & Technical $17,872 $14,649
Managerial $16,352 $16,596
The SWW figures 79 show that males filing alone have much
higher incomes than those of females filing alone. Nine Sts. and
NJ Debtors showed substantially the same difference. The find-
ings of all three studies and CT are tabulated together for com-
parison in Table 6 below.
Female % Of
Males Females Male Incomes
SWW1 2 $15,591 $10,113 64%
Nine Sts. $13,750 $10,427 75%
NJ Debtors 3 $14,300 $10,100 70%
CT, $10,809 $ 9,703 90%
1. SWW, supra note 1, at 151 table 8.1 (includes chapter 13's).
2. Nine Ste., supra note 1, at 306 (roughly weighted).
3. NJ Debtors, aupra note 1, at 550 table 9 (roughly weighted).
4. CT, supra note 1, at 18.
76. SWW, supra note 1, at 152-53.
77. NJ Debtors, supra note 1, at 550 table 9.
78. The numbers are small (males - 104; females - 57).
79. SWW, supra note 1, at 151 table 8.1.
HeinOnline -- 43 Rutgers L. Rev. 207 1990-1991
208 RUTGERS LAW REVIEW [Vol. 43:185
XI. MEDICAL DEBT
Table 7 below compares the equivalent SWW figures with the
consistent information in other recent studies.
Debtors Average of Total
Reporting Amount Unsecured Debt Median
SWW1 51% $1755 11% $616
Nine Sts. 2 56% $1878 12% $567
NJ Debtors3 46% $1420 12% $492
GAO 4 - $ 775 - -
CRC 5 65% $1840 - -
1. SWW, supra note 1, at 168. 1
2. Nine Sts., supra note 1, at 296 table 0.
3. NJ Debtors, supra note 1, at 571 table 33.
4. GAO, supra note 1, at 51.
5. II CRC, supra note 1, at 39.
Another confirmatory finding from their data is that their
chapter 7 debtors owed more medical debt (average = $2032)
than their chapter 13 sample (average = $1154).11 GAO had pre-
viously compared chapter 7's and 13's and had found much the
same difference. The GAO chapter 7's had an average medical
debt of $775 while their 13's were less than half at $360. 81
Medical Debt Scheduled by Chapter
Chapter 7's $2032 $775
Chapter 13's $1154 $360
Chapter 13's as % of
chapter 7's 57% 46%
1. SWW, supra note 1, at 170.
2. GAO, supra note 1, at 51.
In the absence of any information and apparently unaware of
the NJ Debtors' findings, the authors conjecture about the lack of
80. SWW, supra note 1, at 170.
81. GAO, supra note 1, at 51.
HeinOnline -- 43 Rutgers L. Rev. 208 1990-1991
1990] BOOK REVIEW 209
medical insurance.82 In NJ Debtors, 55 of 60 (92%) chapter 7
debtors interviewed later reported having some form of health in-
surance before the date of filing." We had thought that might
account for the lower frequency and smaller amounts of medical
debt than was found in Nine Sts.
The authors' data suggest to them that women filing alone are
afflicted with more medical debt and less income to pay that
debt.84 They do not reference the Nine Sts. and NJ Debtors stud-
ies which had previously shown that women filers have lower in-
comes and larger medical debts. Nine Sts. showed that women's
medical debt as a percentage of their annual incomes is much
higher than that of men alone and joint petitioners. 6 NJ Debtors
showed that women alone have more medical debt 7 and lower
incomes than men filing alone." The authors have replicated the
earlier studies and produced about the same information on med-
ical debt in bankruptcy.
XII. CREDIT CARD DEBT
The authors discursively examine the meaning of credit card
"junkies" to identify those who might be considered "abusers."
Table 10.2, which charts the "Characteristics of Credit Card
Junkies by Three Definitions, '" 9 is an exercise in contemporary
scholasticism with numbers, resulting from the unrestrained use
of computers. Table 9 below shows the important parts of the
SWW findings and compares the SWW replication to the similar
findings in the other studies.
82. SWW, supra note 1, at 171-73.
83. NJ Debtors, supra note 1, at 571 table 34.
84. SWW, supra note 1, at 171.
85. "[A]s a group women debtors earn significantly less .... Women filing alone had
incurred medical expenses ... which constitute a far larger percentage of their incomes
than any of the other categories." Nine Sts., supra note 1, at 289.
86. Id. at 296 table 0.
87. NJ Debtors, supra note 1, at 571 table 33.
88. Id. at 550 table 9.
89. SWW, supra note 1, at 186.
HeinOnline -- 43 Rutgers L. Rev. 209 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
Credit Card Debt
SWW 1 Nine Sts. 2 NJ Debtors3
% of Debtors 88% 71% 88%
Av. $ (for debtors
scheduling) $37414 $20105 Ord. $2537 Ord. $2633 Ord.
$2933 Mdse. $3621 Mdse. $3155 Mdse.
% of total Av. $
unsecured debt for
debtors scheduling 11%6 Ord. 16% Ord. 28% Ord.
credit card debt 37% Mdse. 28% Mdse. 22% Mdse.
SWW, supra note 1, at 183 table 10.1.
Nine Sts., supra note 1, at 297-98 tables Q & R.
NJ Debtors, supra note 1, at 566-68 tables 29 & 30.
Average total of credit card debt. SWW, supra note 1, at 183 table 10.1.
SWW, supra note 1, at 311 table 17.5.
SWW, supra note 1, at 307 table 17.2.
XIII. DEBTORS' PREVIOUS FILINGS
About 8% of all the debtors surveyed in SWW had previously
filed a bankruptcy petition.90 Table 10 below compares the vari-
ous findings which, because of SWW's combining of chapter 7's
and 13's, are not strictly comparable. The GAO, CRC and NJ
Debtors figures apply only to chapter 7 debtors.
NJ Debtors 1%2
Brookings 2% 5
1. SWW, supra note 1, at 192.
2. NJ Debtors, supra note 1, at 586.
3. 1 CRC, supra note 1, at 79. Ten percent for respondents; 11% for non-respondents.
4. GAO, supra note 1, at 22.
5. Brookings, supra note 1, at 59.
6. Chapter 13's.
90. They tell us "these figures are of the same magnitude as those found in the Brook-
ings Study although they are not directly comparable." SWW, supra note 1, at 198 n.7.
The empirical studies from about 1981 are not mentioned.
HeinOnline -- 43 Rutgers L. Rev. 210 1990-1991
1990] BOOK REVIEW
XIV. CHAPTER 13 PLANS
The authors' data on chapter 13 plans are similar to the GAO
findings. The GAO findings go unmentioned, although Sullivan,
Warren and Westbrook do refer to Brookings' study conducted
twenty-five years earlier under a very different bankruptcy law
for chapter 13's. Again, one gets the impression that they want to
clear the stage of all the other actors and be the direct heirs of
Brookings' pioneering study. Table 11 compares the GAO and the
later SWW findings.
Chapter 13 Plans
SWWI(N=591) GAO 2(N=-589)
Average Repayment Proposals 51% 57%
Full Repayment Plans 36% 28%
Zero Repayment Plans (inc. 1% or 3% 4%
1. SWW, supra note 1, at 36, 45 n.27.
2. GAO, supra note 1, at 43, 45.
In asking "Can These Debtors Pay?,"9' 2 SWW provides a helpful
analysis from its large sample of chapter 13 debtors. Most of this
material responds to the credit industry's contention that a sig-
nificant number of chapter 7 debtors could repay some fair por-
tion of their debts in a proceeding like a compulsory chapter 13.
The authors analyze the situations of the sample of debtors and
contend that nine-tenths of the chapter 7 debtors could not pay
anything in a chapter 13. This conclusion is based on unrealistic
assumptions, in particular that the chapter 13 plan would pay
100% to unsecured creditors in three years. 5 Only 36% of the
SWW sample and 28% of the GAO sample even proposed full
payment plans.9 4 The assumptions in their hypothetical chapter
91. Table 11 would likely be different if calculated subsequent to the 1984 amendments
to the Bankruptcy Code. Zero payment plans were widely prohibited before 1984. Before
that time most courts considered the percentage repayment proposed by the debtor in
determining "good faith." The 1984 amendments, 11 U.S.C. § 1325(b)(1)(B) (1988), re-
quire that the debtor apply all his disposable income to the plan if any unsecured creditor
92. SWW, supra note 1, at 199.
93. Id. at 208.
94. See supra Table 11.
HeinOnline -- 43 Rutgers L. Rev. 211 1990-1991
212 RUTGERS LAW REVIEW [Vol. 43:185
13 plans"' are of an unlikely correspondence with any widespread
reality.96 Hence they mean very little.
Their recapitulation' shows in one well-drawn diagram the ap-
parently modest results they claim are to be expected from chap-
ter 13's and the very high attrition rate which resulted in only
about one-third of their debtors still paying on their plans a few
years after filing; another third had definitely failed.es From the
authors we learn nothing of the most important datum, payments
to creditors from the confirmed chapter 13 plans. The whole
point of chapter 13's is to increase the payments to creditors.
95. SWW, supra note 1, at 208-09.
96. In an effort to test how much some homeowner debtors could pay if their homes
were sold, the writers use what they say is a bankruptcy court standard, that the real
estate should bring 70% of its market value at execution sale by auction rather than
through the regular real estate market. They therefore discount the values of the debtors'
homes by 30%. Id. at 204.
The 70% rule of Durrett v. Washington Nat'l Ins. Co., 621 F.2d 201, 203 (5th Cir. 1980)
is the lower limit of allowed real estate execution sale prices that will not be subject to
attack as a fraudulent conveyance. This refers to sales under state law, not in the bank-
ruptcy courts. Others think the bankruptcy courts do better. See LoPucki, A General The-
ory of the Dynamics of the State Remedies/Bankruptcy System, 1982 Wis. L. Rav. 311,
315-21. That is consistent with my experience. The reasons have to do with the great
differences between state court execution sales and the "judicial sales" in bankruptcy
under the control of the bankruptcy court and the trustee. In bankruptcy there is wider
discretion in the terms and conditions and the conduct of the sale, including the notices
and advertising. A real estate broker may be retained. The property can be examined in
advance by prospective buyers. The court can provide that the buyer take title free and
clear of liens or encumbrances. In most states execution sales have little of this flexibility
and there is rarely any guarantee by the sheriff or other auctioneer even that a marketable
title will be delivered.
97. SWW, supra note 1, at 216 figure 12.
98. Claude L. Rice, whose firm provides data processing services for chapter 13's for
various courts, found that about 29.6% of a huge 20,235 case sample of chapter 13's went
through to completion; about 42.5% were dismissed. Future Earnings Hearings, supra
note 33, at 52. The authors' much smaller sample more or less tends to confirm Mr. Rice's
Research in the Western District of New York on a 1971 sample of chapter 13's which
was reexamined in 1976 showed that 27.5% of the original cases in the sample had been
dismissed and 5% adjudicated into straight bankruptcy (liquidation). But 47.5% of the
chapter 13 plans had been successfully completed. Girth, The Bankruptcy Reform Pro-
cess: Maximizing Judicial Control in Wage Earners' Plans, 11 U. MIcH. J.L. Ru. 51, 59
table 1 (1977). These different findings are not mentioned by the writers.
In Wage Earner Plans Under the Bankruptcy Act: Hearings on H.R. 1057 and H.R.
5771 Before Subcomm. No. 4 of the House Comm. on the Judiciary,90th Cong., 1st Sess
48 (1967) [hereinafter 1967 Hearings], it appeared that about 44% of the chapter 13's had
been concluded, another 44% had been dismissed, and the rest were pending.
HeinOnline -- 43 Rutgers L. Rev. 212 1990-1991
1990] BOOK REVIEW
XV. FINANCIAL AND SOCIAL VARIABLES ARE NOT PREDICTIVE
A large part of their materials entitled "Laws, Models, and
Real People""9 is devoted to a restatement of economists' models
of debtor behavior presented at a conference and published in a
symposium on the Economics of Bankruptcy Reform 00 and their
restatement of the critique of that model.
The authors purport to test the economists' model by various
statistical analyses, in particular "by including all the principal
economic variables in a multiple regression equation" to deter-
mine how strong they are as predictors.' 0 ' They decide the eco-
nomic variables have very little or no predictive power as regards
the choice of chapter 7 or 13.102
They comment on psychological matters. Debtors on the verge
of bankruptcy, they say, "are often filled with anxiety and self-
loathing."' 0 8 How do they know that? This is not to cavil; their
book purports to be the report of a scientific study. It is about
facts, they say. But this observation (of whom? of what?) is not
Having rejected the economists' model, the authors examine
the possible impact of various other noneconomic variables. Their
analysis shows a tweak relationship between social-demographic
factors and the decision whether to file in chapter 7 or 13.104
XVI. DIVORCE AND BANKRUPTCY
Sullivan, Warren and Westbrook make passing reference to the
"common finding that divorce is associated with bankruptcy," cit-
ing the 1973 Bankruptcy Commission Report that "[a]ll studies
considering this factor indicated a [divorce] rate higher than for
the general population or for a control group used in the
study."' 0 The authors refer to "[tihe pervasive presence of di-
vorce and separation in [their] files."' 06 Given that information,
they should have designed their study to examine this factor as
99. SWW, supra note 1, at 230.
100. The Economics of Bankruptcy Reform, 41 LAw & CorrEP. PROBS. 1 (Autumn
101. SWW, supra note 1, at 235.
102. Id. at 236-37.
103. Id. at 244.
104. Id. at 246. See also id. at 236.
105. Id. at 160-61 n.8.
106. Id. at 296.
HeinOnline -- 43 Rutgers L. Rev. 213 1990-1991
214 RUTGERS LAW REVIEW [Vol. 43:185
possibly contributing to the 1980's increase in personal bank-
ruptcy filings and choice of chapter. Also, divorce is an area in
which they say there is that paucity of information which caused
them to begin their study and gather information on, among
other things, the marital status of bankruptcy debtors.
The only information the writers had relating to divorce was
from local rule "interrogatories" from chapter 7's in the Western
District of Texas. Very few of the seventy-five chapter 7 debtor
sample (3.4%) reported that divorce and related family problems
were the cause of bankruptcy.' The authors did not know
whether the chapter 13's in the Western District of Texas were
similar to the chapter 7's in this regard as was the case for family
size. l09 But the chapter 13 filings provide enough information so
that most recent or pending divorces could be discovered." 0
Given the several consistent studies over time and the available
data on rates of divorce and personal bankruptcy, one supposes
this would be a key social (i.e., noneconomic) variable to be ex-
amined."' But in testing marital status for its power to predict
debtors' choice of chapter 7 or 13, they apparently recorded only
married or single status and found but a weak correlation be-
tween joint (i.e., married debtors) filings and choice of chapter
XVII. DEFICIENCIES OF THE ECONOMIC MODEL
Their data suggest to the authors two discouraging conclusions:
107. See id. at 342.
108. Id. at 161 n.8. The causal relationship, if any, could have gone in the other
109. Id. at 81 n.13. See also supra Part VII.A.
110. Official Form No. 10 for chapter 13 asks about separation, alimony, maintenance or
support payments, dependents and codebtors (usually a spouse). Question 4 calls for a
detailed family budget which might also reveal information on divorce.
111. In CRC's respondent sample, 20% were divorced and 9% were separated at the
time of filing. 11 CRC, supra note 1, at 6 exhibit 1-3.
GAO's analysis found that bankruptcy and divorce correlated with a moderately strong
7.2 coefficient. GAO, supra note 1, at 17-18.
112. SWW, supra note 1, at 245. The 1973 Bankruptcy Commission Report would have
been correct in its statement of correlation at least until the mid 1980's because the rates
of bankruptcy and divorce appeared reasonably well-correlated during those years. Al-
though the national divorce rate has leveled off since 1980 (5.2 per thousand persons in
1980 to 4.8 per thousand in 1986), the frequency of personal bankruptcies continued to
rise during most of those seven years. This is yet another example of why time series
studies are so important in establishing the external validity of a (perhaps causal) relation-
ship between bankruptcy and any other variable.
HeinOnline -- 43 Rutgers L. Rev. 214 1990-1991
1990] BOOK REVIEW
(1) the simple economic model is worthless as a predictor of
debtor behavior; and (2) the relative importance of noneconomic
factors is unknown."'
They relate that "[slome bankruptcy experts began to evaluate
proposed changes in the bankruptcy laws based on the 'incen-
tives' these laws created either for debtors to stay out of bank-
ruptcy or for debtors to choose Chapter XIII repayments rather
than Chapter VII discharges.""" SWW's references to the "intel-
lectual movement" before the 1978 Code urging incentives for
debtors to select chapter 13 rather than Chapter 7 are hardly con-
vincing.115 "Other commentators," they say, "carried the idea."
Those others are two law students.1 They speak of "commenta-
tors" discussing the possibilities that minimum federal exemp-
tions would increase the number of filings but that differences in
state exemptions would cause forum shopping. ' The support for
this assertion is that three employee-spokesmen of the consumer
credit industry made those claims.' 8
In the SWW statement of the premises of the economic or in-
centive model, two are worth examination: "[The] disincentives to
use bankruptcy and incentives to choose Chapter 13 payouts will
strongly influence debtors to pay their debts outside of bank-
ruptcy or in Chanter 13. ' '11 This is a loaded proposition. Whether
there will be a strong influence or any perceptible effect at all
depends not only on the predictive power of the model but also
113. SWW, supra note 1, at 254.
114. Id. at 232. The choice of chapter 7 or 13 was not an issue at the Conference on the
Economics of Bankruptcy Reform which emphasized the question of incentives to file and
the effects upon the consumer credit market.
115. One gets the impression that after 1979 many scholars were urging greater use of
chapter 13. ("Chapter 13 is popular among policy-makers and academics alike." Id. at
223.) In fact, so far as I can tell, a few economists and law teachers were saying something
defensive: If the recent increases in chapter 7 filings are disturbing to the Congress (the
more so because of the consumer credit industry's intense lobbying for a compulsory chap-
ter 13), one possible remedy is legislation to make chapter 7 more expensive and chapter
13 more attractive. Others were merely noting or opining that the new Code had resulted
in more filings, but most agreed that the new law did not account for the entire increase
from 1979 through 1982.
There was an article in favor of a system in which discharges would be conditioned on
the debtor's future repayment efforts. See Eisenberg, Bankruptcy Law in Perspective, 28
UCLA L. Rzv. 953, 978-91 (1981).
116. See SWW, supra note 1, at 264 n.5.
117. Id. at 232.
118. See id. at 265 n.9.
119. Id. at 234.
HeinOnline -- 43 Rutgers L. Rev. 215 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
on the strength of the incentives and penalties. SWW overlooks
the difference between explanation by means of a model, and pre-
diction. The economic model may explain a large part of relevant
debtor behavior. But it may predict that behavior well only near
the margin, when the differences-the incentives-are larger than
what we have even in these three states.
The second premise is "that property exemptions are the key
incentive/disincentive in influencing debtor behavior."120 There is
ample evidence that the exemptions from wage garnishment are
more important. 21 But property exemptions, although not an in-
dicator of the frequency of bankruptcy filings,122 are one of the
potentially decisive factors.
To test their version of the economic model, the authors as-
sume that property exemptions will directly influence some debt-
ors who could repay something1 23
to: (1) avoid bankruptcy, or (2)
choose chapter 13. They say they will test each prediction of the
economic model to see if exemption levels influence the decision
to file or the choice of chapter 7 or 13 among those who ulti-
121. See supra Part V.
122. See Woodward & Woodward, Exemptions as an Incentive to Voluntary Bank-
ruptcy: An Empirical Study, 88 CoM. L.J. 309 (June/July 1983), reprinted in 57 AM.
BANKR. L.J. 53 (1983). GAO found that "States within both groups--those opting out and
those not opting out-experienced a wide range of increases and/or decreases which indi-
cates that factors other than exemptions may have influenced an individual's decision to
file bankruptcy." GAO, supra note 1, at 28. This is supported in a series of state by state
tables of those states that did and did not opt out (the former with smaller exemptions)
which show no consistent pattern of increases or decreases in the frequencies of chapter 7
and chapter 13 filings. Id. at 83-88 appendix 11. CT had preliminarily noted the lack of
any consistent relationship between frequency of filings and smaller exemptions (states
that had opted out) or larger exemptions (states that still retained the federal exemp-
tions). CT, supra note 1, at 3 tables A-B.
123. It is not clear what they mean by "can pay" debtors. SWW, supra note 1, at 235.
The ability to pay can come from income. Hence debt to income ratio is a variable to be
examined. The ability to pay can also come from the sale of assets. Most of the persons
using such a model refer to assets. This is evidently what Professor Vukowich meant when
he spoke of reduced bankruptcy exemptions as a disincentive resulting in fewer filings. His
examples are debtors who "surrender some valued assets," debtors who "retain significant
wealth," and "debtors with nonexempt assets [who] would have a stronger incentive to file
in chapter 13, which permits debtors to retain all their property." Vukowich, Reforming
the Bankruptcy Reform Act of 1978: An Alternative Approach, 71 Gzo. L.J. 1129, 1152
Similarly, CRC concludes that "consumers with few assets and great quantities of un-
secured debts have much to gain by declaring bankruptcy. Conversely, those with a great
amount of assets relative to liabilities have little to gain by declaring bankruptcy." II
supra note 1, at 142.
HeinOnline -- 43 Rutgers L. Rev. 216 1990-1991
1990] BOOK REVIEW
mately do file in bankruptcy.1 2 '
It follows from their view, if the economic model is correct,
high exemptions should result in more debtors in bankruptcy
with greater capacities to pay; low exemptions should deter all
but those debtors least able to pay. Similarly, their data should
show that high exemptions encourage debtors to use chapter 7,
while low exemptions should force all but the most impoverished
debtors into chapter 13. Their test of the economic model looks to
the debtors' ability to pay as measured by incomes.
The authors say their data show that all the correlations that
would support what they perceive as the predictions of the eco-
nomic model are simply too weak to establish anything."" In par-
ticular, (1) the debt-income ratios of debtors in chapters 7 and 13
are about the same' 2 ' and (2) the debt-income ratios in Texas,
Pennsylvania and Illinois (with differing property exemptions)
are effectively the same. 2 7 These weak or negative correlations
are true of their sample as calculated both with and without
mortgage debt. They conclude that "these direct tests of the eco-
nomic model demonstrate that the model. . . does not accurately
predict debtor behavior.' ' 28 They add that their "direct test,"
showing that the economic model does not predict, should affect
not only bankruptcy legislation, but also "because other legal pol-
icies have been affected by similar economic models, [their] find-
ings in bankruptcy should affect how those models are used else-
124. SWW, supra note 1, at 235.
125. Some of the authors' own data, however, appear to bear out the predictions of this
economic/incentives model which they think is worthless. For example, most of the debt-
ors with higher net worth are in chapter 13 "trying to pay debts and keep property." Id. at
73. Their sample of Texas debtors (with a much larger property exemption available to
them) who are balance sheet solvent "had a net worth three times higher than the average
in Illinois or Pennsylvania." Id. These data also tend to fulfill the predictions of the eco-
nomic/incentives model that higher exemptions will result in wealthier persons (i.e., those
with more assets) filing in bankruptcy.
Florida provides another example. The State has very generous exemptions which
would, in the jargon used, make it more attractive for personal bankruptcy filings. Since
1980, personal bankruptcy filings in Florida have increased at more than three times the
national rate. Brannigan, Floridais Gaining Reputation as Haven For Debtors Who Seek
to Shelter Wealth, Wall St. J., Aug. 3, 1990, at B1.
126. Id. at 239.
127. Id. at 241. These are not new findings. GAO's study in New York, Ohio, Kentucky
and California had found that "chapters 7 and 13 debtors have similar income to debt
relationships." GAO, supra note 1, at 73, 53.
128. SWW, supra note 1 at 242.
HeinOnline -- 43 Rutgers L. Rev. 217 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
where."' 1 9 These are quite the boldest claims in a book replete
with grandiose claims for its supposed accomplishments.
As I indicated earlier,130 bankruptcy rates in various states ap-
peared to be related to the severity of their garnishment laws. 3 1
To finesse this variable as they do is to create an unnecessary
problem, especially as it appears that wage garnishment laws are
a more important correlate of bankruptcy than are property ex-
emption laws.132 This is a crucial lack, given their choice of states.
Texas and Pennsylvania had no wage garnishment. 83 Another
problem is the mistaken results from their using debt figures as of
the filing date and relating them to incomes for the preceding cal-
endar year. Those incomes will be lower than incomes for the year
of filing, while the debts are apt to be lower at the date of filing
than they were the year before.
The home is usually the largest and most important investment
for most persons and families. If a debtor's homestead equity is
less than the allowable state exemption, my economic model
would predict that most such debtors are not apt to file in bank-
ruptcy unless there are other compelling reasons. Only some frac-
tion of such homeowner debtors will file at all in a state like
Texas, with an unlimited homestead exemption (and no wage gar-
nishment). Nor can the writers, from their data, test the proposi-
tion that homestead and other property exemptions influence
debtors' decisions to file, because they have not tested this rela-
tionship over time and they have no sample of and no data on
non-filer debtors in similar financial situations.
The authors' analysis of homeownership and the choice of
chapter 7 or 13 is likely to be based on an odd sample of those
debtors remaining after the loss of the larger set of potential
bankruptcy debtors who do not file. They have not tested the ec-
onomic model at all for choice of filing, and they have not tested
directly for choice of chapter 7 or 13.
129. Id. at 10.
130. See supra Part V.
131. Brookings, supra note 1, at 28-32, 49. See also Peterson & Aoki, supra note 29, at
101 n.4 (listing similar studies).
132. Peterson & Aoki, supra note 29, at 100, 103, 105.
133. Sullivan, Warren and Westbrook did test whether repossessions and property
seizures were predictors of the decision to select chapter 7 or 13, SWW, supra note 1, at
269 n.37, but they performed no similar test for wage garnishment, evidently a more im-
portant variable in relation to bankruptcy. Presumably, this is because they had few such
events due to their unfortunate selection of states.
HeinOnline -- 43 Rutgers L. Rev. 218 1990-1991
1990] BOOK REVIEW
The apparent conclusions made by the writers that property
exemptions have little to do with (1) the decision to file, and (2)
the choice of chapter 7 or 13 may not be warranted on the evi-
dence they present.1 3 ' It had been shown that existing property
exemptions did not predict frequency of filings. 138 They now add
that the same variable does not predict choice of chapter, proba-
bly a less important decision to debtors. Yet they claim the latter
lack of correlation is a direct test of the economic model while,
presumably, the former is not.
Sullivan, Warren and Westbrook have not examined the first
question, whether much-increased federal exemptions result in
more filings or possibly fewer filings if the larger exemptions are
under state law. Whether larger property exemptions could make
much difference in the decision to file, and in which chapter, will
depend on how much and what types of property-exempt or
not-are owned by a set of potential debtors, and maybe on
whether the exemptions are state-created and therefore applica-
ble independently of and prior to bankruptcy.
The differences between property exemption laws are often dif-
ficult to quantify. For example, state homestead exemptions can-
not be "spilled over" to exempt other property of the debtor if
they are not used, or used only in part to exempt the debtor's
homestead. The federal exemption law allowed a full spillover of
up to $7,900 per debtor in 1981.3' This flexibility made the fed-
eral homestead exemption worth more in the sense that it was
much more apt to be used by a set of chapter 7 debtors, only half
of which (if we believe the authors), or 16% to 28% if we follow
the other empirical studies, 1 7 were homeowners. I think few of
the homeowner debtors had equity in 1981 that exceeded the fed-
eral homestead exemption.
It may also be that the economists' model is correct but that
134. Id. at 260 table 13.7. This table gives an average value of $14,498 for exempt prop-
erty. Whether this was scheduled or actually claimed is not clear. It may not matter for
SWW's purposes so long as that was the amount of the debtors' exempt property. The
authors later refer to "debtor's exemption claims." Id. at 237.
The information in table 13.7 is taken from 1469 cases, which means that at least 531
chapter 13's were used for this calculation. Since the amount of the exempt property is a
variable which may have to do with the choice of chapter 7 or 13, combining the two for
this purpose may make such a test very tenuous.
135. See supra note 122.
136. SWW provides no data on the use of spillover.
137. See supra Table 1.
HeinOnline -- 43 Rutgers L. Rev. 219 1990-1991
220 RUTGERS LAW REVIEW [Vol. 43:185
the curve showing increases in filings only begins to rise and be-
come predictive when the difference between the federal and
state exemptions is greater in amount, and the state exemptions
remain much less useful. Posit a state in which less than $1000 of
any type of property is exempt. If the state has not opted out, the
federal exemption will be selected by more debtors. 88 I suppose if
this state had opted out, the very small exemption without spill-
over would result in a greater frequency of chapter 13 elections.
These assumptions are as realistic as concluding from data on
three states that all property exemptions have no influence, or at
least no predictive power, on filing and choice of chapter. It
would be prudent and more accurate to say that within the prop-
erty exemption ranges of their three states, the arguably causal
factor of debt to income ratio appeared not to predict choice of
chapter 7 or 13. This also presents another reason for choosing
more states with greater variations in exemption laws.
Although the authors refer to "The Spherical Chicken,"' 9 I
think they misunderstand its import. The purpose of that article
was not to argue "the lack of empirical verification" of the eco-
nomic model, as they believe. 4 ' That does not invalidate a the-
ory. It is an attempt to show that there are too many rebutting
instances to the general laws which the economists claim that
their models describe, a more formidable objection.
One can, of course, as the writers have done, claim that a rebut-
ting instance (a widespread result which would not be predicted
by the model) is a direct and dispositive test of the economic
model. They fail in this undertaking because property exemptions
within this range are not predictive even of filing and because
their data on debt-income ratios do not present rebutting in-
stances. Even were the debt to income ratio a direct test, the cal-
culations based on their data are inaccurate and not apt to be
probative because the incomes are for the preceding calendar year
and will be significantly lower than present (year of filing) in-
comes of their sample of debtors. The debts, however, are as of
138. Apart from the economic model, that much seems clear from their Pennsylvania
data, where only 6% of their debtors took the Pennsylvania state exemptions. SWW,
supra note 1, at 43 n.19. In the similar New Jersey situation (New Jersey, with small state
exemptions, had not opted out), none of the 186-debtor sample chose the state exemption.
NJ Debtors, supra note 1, at 553 n.33.
139. Shuchman, Theory and Reality in Bankruptcy: The Spherical Chicken, 41 LAW &
CONTEMP. PROBs. 66 (Autumn 1977).
140. SWW, supra note 1, at 264 n.8.
HeinOnline -- 43 Rutgers L. Rev. 220 1990-1991
1990] BOOK REVIEW
the filing date and may have been reduced over that time.
XVIII. PERSUADING DEBTORS TO SELECT CHAPTER 13
The authors have overlooked much recent published history in
their discussion of why debtors selected chapter 13, or they exag-
gerate to the point of hyperbole. They speak of "Another Possible
Influence-Local Legal Cultures" and state that they
believe that debtors' bankruptcy decisions may be importantly
influenced by the local legal culture they encounter when they
seek legal help. We believe that the "uniform" system of bank-
ruptcy laws is in fact very different in different courts.. . . The
clue that there were significant influences at the local level came
initially from the reports published by the Administrative Office
of the Courts indicating .. .what proportions of . . .debtors
file in Chapter 13. . . . [Ilt is clear that there are huge differ-
ences in the proportions of Chapter 7's and 13's filed in different
districts. From the very start of our study, we suspected that
these district variations must contain an important element of
the explanation for debtors' choice of chapters.'
They then go on for several pages to relate their detective story
about trying to cu4l an explanation from their data. They are not
quite able to figure it out but they tentatively conclude "that at-
torneys may exert a powerful influence over choice of chapter."""2
It is also true that bankruptcy judges influence the behavior of
lawyers. However, no clues or suspicions were needed. It has been
fairly common knowledge for more than thirty years that there
are dramatic differences in the chapter 13 filing rates in several
districts. Not only have the A.O. figures revealed those districts
every year, but the great differences, including much of the his-
tory and some of the practices, were related to the Congress in
1967,14" and before that in the journal literature. " Both these
publications contain discussions of the history of chapter 13 fil-
ings and how judges influence both lawyers and debtors in their
choice of chapter 13 and obtain the cooperation of creditors in the
Sullivan, Warren and Westbrook should have been aware of the
141. Id. at 246.
142. Id. at 251.
143. 1967 Hearings, supra note 98.
144. See McDuffee, The Wage Earner's Plan in Practice, 15 VAND. L. REv. 173 (1961).
HeinOnline -- 43 Rutgers L. Rev. 221 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
considerable commentary on these long-standing patterns of the
differing frequency of chapter 13 use." ' Brookings devoted a sec-
tion to the influence of the bankruptcy judges and the lawyers on
the debtors' choice. 1"6 Jacob14 7 found a great disparity in the use
of chapter 13's in Wisconsin. "The preponderance of Chapter 13
cases from Madison is no sampling accident," Jacob explained.
The Referee was a "strong advocate of Chapter 13; he communi-
cated his enthusiasm to the bar and assisted lawyers in filing
Chapter 13['s]"; he urged lawyers to use chapter 13; he provided a
good chapter 13 trustee to help the lawyers.
Informal coercion is not apt to account for most of the 93,000
chapter 13 filings in 1981 nor for the 183,000 in 1989. The
debtor's being able to keep the real or personal property collat-
eral for a secured loan accounts for many of the chapter 13 filings.
Also, the authors are incorrect in stating that "[tihe source of
variation in bankruptcy had always been assumed to be the
states.... The reasoning was that the exemptions are affected
by state law." 49 This statement is called for by their critique of
the economic model. The model is supposed by the authors to be
based solely on the motivations created by varying property ex-
emptions. Therefore, there should be no difference between sets
of bankruptcy debtbrs in different districts within a state because
the exemption law is the same for all of them. But this had been
shown not to be always true. The 1977 "Spherical Chicken" arti-
cle made the point that there is great intrastate variation which a
narrowly construed economic model would not predict.
New York State provides [an] example. It has four federal dis-
tricts. In three of those districts the rate of XII's ranges from
one percent to two percent of all personal employee filings. In
the fourth district about 29% of all employee filings are for
XIII. . . . Here the applicable state law is the same.150
145. More discussion may be found in the H.R. Doc. No. 137, 93d Cong., Ist Sess., pt.1,
at 157-58 (1973) (Report of the Commission on the Bankruptcy Laws of the United
States), S. REP. No. 989, 95th Cong., 2d Sess. 12 (1978), and H.R. REP. No. 595, 95th Cong.,
1st Sess. 117 (1977). See also Shuchman, supra note 139, at 84-85.
146. Brookings, supra note 1, at 74-76.
147. H. JACOB, DEBTORS IN COURT 67 table IV-11 (1969).
148. Id. at 66.
149. SWW, supra note 1, at 343.
150. Shuchman, supra note 139, at 85.
HeinOnline -- 43 Rutgers L. Rev. 222 1990-1991
1990] BOOK RE VIEW 223
Chapter XIII Filings as Percentage of all
Personal (Nonbusiness) Bankruptcies]
1970 1971 1972 1973 1974 1975
Total Nonbusiness 1,851 2,092 1,869 1,571 1,711 2,481
Chapter XIII 4 14 11 8 10 19
% 0.2% 0.7%. 0.6% 0.5% 0.6% 0.8%
Total Nonbusiness 617 921 957 884 1,094 1,991
Chapter XIII 3 11 11 5 9 26
% 0.5% 1.2% 1.1% 0.6% 0.8% 1.3%
Total Nonbusiness 446 718 816 703 875 1,354
Chapter XIII 3 6 4 1 12 26
% 0.7% 0.8% 0.5% 0.1% 1.4% 1.9%
Total Nonbusiness 1,999 2,423 2,393 2,304 2,639 4,199
Chapter XIII 52 280 303 369 546 1,196
% 2.6% 11.6% 12.7% 16.0% 20.7% 28.5%
1. Shuchman, supra note 139, at 85.
Professor Marjorie Girth later explained that wage earners'
plans had rarely been used in the Western District of New York.
The newly appointed bankruptcy judge came to strongly favor
the use of chapter 13's. "He launched an educational effort aimed
at the practicing bar and at creditors, appointed a standing trus-
tee for all such cases, and made arrangements to utilize a comput-
erized service for maintaining case records and handling disburse-
ments.""' Professor Girth showed that the number of chapter
13's had increased by twenty-one times from 1970 to 1975.1" The
disbursements to creditors increased enormously in those six
years, from $54 to some $1.2 million. 58
XIX. RELUCTANT CREDITORS
In the second part of their book, the authors discuss what they
claim is their pioneering research on the creditors commonly
151. Girth, supra note 98, at 56.
152. Id. at 56 n.36.
HeinOnline -- 43 Rutgers L. Rev. 223 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
listed in the bankruptcy schedules. "We also decided to collect
data about the creditors of the debtors in bankruptcy. No one
had ever done that before. That no one had investigated creditors
before may tell us something interesting about unconscious atti-
tudes toward the bankruptcy process. ' "
They comment on what they term "reluctant creditors," that
"[t]he mere fact of reporting the position of reluctant creditors
for the first time creates a new perspective. For years people have
talked vaguely about medical losses or tort debt in bankruptcy,
with no clear idea what was at stake.' '1 "
I do not know how their information, replicating and largely
confirming earlier studies, could create a new perspective."" They
include in this category five common types of debt: educational
loans, medical debt, taxes, rent and utilities. These are docu-
mented below in Table 13, which also compares their findings
with those of Nine Sts. and NJ Debtors.
XX. CONSUMER CREDITORS
The authors begin their discussion with a counterpoint. Having
put to the reader how knowledgeable and aware are the firms in
the business of consumer credit of the risks of default, one as-
sumes, they say, that these firms will be insulated from large
bankruptcy losses. How wrong is that reader. In fact, "[t]he
consumer credit industry is the biggest loser in consumer bank-
ruptcy." 1 "Our data show that the most bankruptcy losses are
not borne by inexperienced lenders or by reluctant creditors.
Measured in absolute dollars, consumer bankruptcy losses are lost
by professionals, those who actively sought consumer lending for
154. SWW, supra note 1, at 342.
155. Id. at 299.
156. "Reluctant creditors" is yet another illustration of Maier's Law. Old findings are
rediscovered. But there is no mention of their earlier existence. On the contrary, we are
told that "no one had ever done that before," and that the mere fact of reporting these
matters "creates a new perspective." Then the rediscovery is given a new name; with that
new identity comes a new life, the reincarnation.
157. "Only the occasional clever debtor or the lucky beneficiary of a bizarre computer
blip might slip through the consumer credit industry's checks and balances and thus tum-
ble into bankruptcy owing a substantial amount to this well-run machine." SWW, supra
note 1, at 303.
159. Id. at 304.
HeinOnline -- 43 Rutgers L. Rev. 224 1990-1991
1990] BOOK REVIEW 225
69i 69% A
.- a _ . .
609 GOD, 603
EnZ 5 Cq
*- .i 4 %6 ...........-..
HeinOnline -- 43 Rutgers L. Rev. 225 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
Perhaps the authors merely feign their air of surprise. Perhaps
they think readers will be surprised that those in the business of
lending money and who lend the most money to individuals also
have the largest losses in bankruptcy.1e0 Those facts have been
evident in every empirical study of bankruptcy. Other researchers
took it for granted that the consumer credit business firms would
provide most of the credit to consumers"' and, in the event of
bankruptcy, would bear the largest part of the losses.
A. Undersecured Creditors
The authors claim to have "taken a first step in introducing
facts about creditors and their lending practices into the policy
debates."' 62 That is an extraordinary statement even for this
book, and it is untrue. There is an enormous and well-known
literature on the law, practices, and finances of the consumer
credit industry in its various forms.'""
They also claim knowledge of a new datum: "We are able to
estimate for the first time the actual exposure of the [consumer
credit] industry as opposed to its scheduled bankruptcy debts."'"
This is hardly new information, nor is it estimated for the first
time in this book. It appears to them that many secured creditors
are undersecured it personal bankruptcies and that they may lose
as much as 10% of their claims.
Secured creditors are the least affected in bankruptcy and, in
terms of the impact of bankruptcy upon them, are much less im-
portant than general creditors. The largest fully secured creditor
in the nation, General Motors Acceptance Corporation
("GMAC") wrote off 0.42% of its consumer credit receivables in
1981 due to all defaults including bankruptcy.'1 5 Hence the usual
160. "The most important type [of creditors] both by frequency and number were firms
in the business of extending unsecured consumer credit that, one would suppose, were best
able to gauge risks, and could also charge the most for money." NJ Debtors, supra note 1,
161. Early on, H. JACOB, supra note 147, at 63 table IV-8, presented a table of the larg-
est credit sources of bankrupts and "other delinquents." It showed that the largest credit
sources to bankrupts were consumer credit business firms, banks, and finance companies.
The latter two lenders accounted for 77% of the bankrupt's credit, 21% and 56%,
162. SWW, supra note 1, at 322.
163. See infra notes 207-09 and accompanying text.
164. SWW, supra note 1, at 302.
165. Personal Bankruptcy, Oversight Hearings Before the Subcomm. on Monopolies
and Commercial Law of the House Comm. on the Judiciary,97th Cong., 1st & 2d Seas.
HeinOnline -- 43 Rutgers L. Rev. 226 1990-1991
1990] BOOK REVIEW
emphasis of most of the empirical studies and most of the con-
gressional hearings is on the general creditors and the debtors.
Their table 17.3 shows the unsecured portions of secured
loans.' The authors do not know whether the creditor was un-
dersecured when the original loan was made or whether there was
a refinancing. The undersecurity is at bankruptcy, given the valu-
ation of the collateral by the debtor and the balance of the se-
cured debt also given by the debtor.
For many secured loans there is reaffirmation, because of the
opportunity costs of replacing a car, for example. Secured cred-
itors can get their collateral or its value at bankruptcy. There
may be fluctuations in value or depreciation so that the loan be-
comes undersecured, but secured creditors do not experience the
discharge of their entire loan balance as do most general
They compare the debtor's statement of the value of the collat-
eral at the time of bankruptcy with the loan balance. "[Tiable
17.3 shows that the secured claims, on average, were backed up
by collateral worth at least 88% of the outstanding loan. And a
debtor will often pay 100% of the value to keep the property,"'
thus reducing the bankruptcy risk even more.
In 1982, the Qredit Research Center published its report on
consumer bankruptcy. Since CRC was largely financed by the
consumer credit industry, I suppose the authors were aware that
the consumer credit firms knew that more than half of their most
common secured loans were, in the event of bankruptcy, under-
secured to some extent. "Secured credit. . . was most frequently
secured by personal vehicles and household furniture. For more
than half of these credit obligations . . . creditors had considera-
ble risk exposure in that the outstanding balance was greater
than the market value stated on the petition.""'
565 (1981-82) [hereinafter 1981-82 Hearings].
166. SWW, supra note 1, at 308.
167. "About 20% of the wage-earning debtors in Chapter 7 reaffirmed one or more
debts . . . from the secured column. . . [W]e found considerable evidence of informal
reaffirmations . . . and suspect they were at least as prevalent as formal ones." SWW,
supra note 1, at 319.
168. Id. The writers may mean that a debtor will often pay the full outstanding loan
balance even if the value of the collateral is less, or they may be referring to the debtor's
redemption under Code § 722, 11 U.S.C. § 722 (1988), by cash payment of the fair market
value of the property up to the loan balance.
169. II CRC, supra note 1, at 32-34.
HeinOnline -- 43 Rutgers L. Rev. 227 1990-1991
228 RUTGERS LAW REVIEW [Vol. 43:185
CRC had found "that average secured debt holdings on bank-
ruptcy petitions is less than the market value of property used to
secure the debt. '17
In gross, aggregated terms, the average differ-
ence is small, about 2% of the debt, or $314, in more than a thou-
sand chapter 7 bankruptcies. 171 In CRC's set of nonrespondents,
the creditors are oversecured on average only about $10.1'
CRC cast these data as undersecured loans rather than as un-
dersecured creditors. '7 The CRC format, by type of security, is
probably better because different types of collateral depreciate at
very different rates, and are more or less easily replaceable by the
debtor. The most common types of collateral, cars and household
furniture and appliances, are especially different in these regards
as well as in cost and ease of resale.
Since most types of creditors sell secured money for various
types of c6llateral, any advice from the authors would be better
addressed to the kind of collateral rather than to a particular
type of creditor. Perhaps only the captive financers of the auto
manufacturers restrict themselves to motor vehicles. But they an-
alyze their losses very carefully, lobby very effectively and need
no advice from law teachers.
Nonmortgage Secured Credit
Ratio of Loan Balance
Collateral % of Debtors Average $ to Market Value
Cars 50% $3,719 more than 100% = 45%
less than 100% = 55%
Furniture & 41% $ 859 more than 100% = 39%
Appliances less than 100% = 61%
The writers should have compared their data to that in CRC.
They knew the type of collateral from the schedules and,
whatever else they wanted to present, could have given readers
and legislators a better basis for evaluating the CRC data and
The authors make some very bold and, so far as I can tell, un-
170. I CRC, supra note 1, at 69.
171. See id. at 79.
172. See id.
174. II CRC, supra note 1, at 33.
HeinOnline -- 43 Rutgers L. Rev. 228 1990-1991
1990] BOOK REVIEW
supported statements from their data. For example, they say:
"These lenders made secured loans knowing that in every case
the collateral would be worth less than the debt and that the loan
would be 'under water' at some point in the life of the loan."' 7 5
Since Sullivan, Warren and Westbrook do not know the price
of the collateral when it was bought and financed, and since they
do not tell us the amount of the original loan, if they know
that,1 their assertion, on the kindest interpretation, is a guess.
They have no basis for defaming these lenders.177 What the au-
thors do tell us is that automobile finance companies are owed
more at bankruptcy than the value of their collateral, that they
are about 29% unsecured.1
B. Conclusions About Consumer Creditors
The authors' conclusions about consumer creditors are not
based on their data and will come as no surprise. First they con-
clude that "some creditors find there are profits to be made in
unsecured loans that outweigh the risk of bankruptcy losses "17'
175. SWW, supra note 1,at 322.
176. Schedule A-2 of Official Form No. 6 might provide some of that information. Col-
umn 3 calls for specifying when the secured claim was incurred and the consideration
The authors point out that "[tihe forms require notation of the date the debt was in-
curred, but most lawyers do not report the data." SWW, supra note 1, at 326 n.12. Thus
they may know very little of the original transaction and nothing of refinancings, debt
Dr. Sullivan's letter states that they "coded both the market value of the collateral and
the claim value if it were (sic] different. The debt was listed as the attorney listed it
Letter, supra note 44.
177. Even were this "under water" assertion true of cars, in the example they give, it
may very well be a prudent business decision for the manufacturer's finance company to
finance more of the purchase price (or all if the selling price is high enough), and on terms
by which the value of the car will decrease faster than the balance of the loan. The car gets
sold and there may be more profit from the sale and the financing than there is estimated
or actual loss in the event of default or bankruptcy.
This is even more true for the other common types of collateral. "Retail markups range
as high as 100% for furniture and other consumer goods which are often the subject of
security interests." Whitford, The Appropriate Role of Security Interests in Consumer
Transactions,7 CARDoZo L. Rav. 959, 961 (1986) (footnote omitted).
178. SWW, supra note 1, at 308 table 17.3. NJ Debtors had reported that cars, by far
the largest personal property collateral, were on average slightly undersecured by less than
2% (average balance = $3752; average market value = $3688; difference = $64). NJ
Debtors, supra note 1, at 578. Nine Sts., a larger sample, showed an average undersecurity
of about 22% (average balance = $4266; average market value = $3329; difference =
$937). Nine Sts., supra note 1, at 304. The authors do not mention these findings.
179. SWW, supra note 1, at 312.
HeinOnline -- 43 Rutgers L. Rev. 229 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
and, second, I suppose from general microeconomic principles,
that "[s]o long as the overall losses can be held low enough to
permit a reasonable profit on the lending activity, the creditor
can continue to lend .. . ."18O It is silly to imply that these con-
clusions are induced from their findings.
The authors' advice to the consumer credit industry: Don't lend
to persons with debt/income ratios much greater than 0.2%."'
The keys to reducing bankruptcy losses are to avoid such borrow-
ers and also to avoid those who have had "significant income in-
terruptions," i.e., the unemployed."' Having indicated that so
much of the secured debt-perhaps 90%-is likely to be re-
paid, 68 after several pages of rather discursive discussion, the au-
thors conclude that by following their advice and using these two
criteria, "the actual benefits to creditors may not be sufficiently
great to persuade the credit industry to change its current prac-
tices."1 " They speculate, based on their data, that the total loss
for 300,000 wage earner bankruptcies a year would be about $2
billion in 1981, or about 0.05% of the then outstanding $389 bil-
lion in consumer debt. Therefore, they conclude that "the con-
sumer bankruptcy losses may not be enough to make reformed
credit practices seem worthwhile. " 85 Their method seems a diffi-
cult and roundabotit means for approximating this information.
Other sources include the annual reports of the publicly held con-
sumer credit firms and their bond prospectuses and Securities
and Exchange Commission filings which contain reliable informa-
181. "Our data show that what most clearly distinguishes debtors in bankruptcy from
the rest of the population is their ratio of debts to income. While 95% of the population
has a debt/income ratio of less than 0.2, more than 90% of the debtors in bankruptcy had
an equivalent debt/income ratio in excess of 0.2." Id. at 313.
These data demonstrate a one-time relationship during recession years, 1980 and 1981.
The writers do not know whether even these extreme figures would hold true over time or,
for example, during periods of economic growth. Moreover, this 5% of the general popula-
tion is comprised mainly of persons who do not file in bankruptcy. The authors should be
more careful. If their caution to consumer creditors is to mean anything, it should state
that, of those persons with debt to income ratios of 0.2% or more, a fraction of that set
will file in bankruptcy. The size of that fraction would help inform consumer creditors and
readers. Maybe the 0.2% figure demarks a group at high risk for bankruptcy. The size of
the fraction might also reveal that the risk is slight.
182. Id. at 315.
183. Id. at 304.
184. Id. at 320.
185. Id. These statements sound very much like Luckett's conclusion. See Luckett,
supra note 67, at 599. See also infra text accompanying note 194.
HeinOnline -- 43 Rutgers L. Rev. 230 1990-1991
1990] BOOK REVIEW
tion on net writeoffs of consumer credit from all sources, includ-
ing bankruptcy. Such figures were presented in the 1982 Hearings
for several years for several large consumer creditor firms."'
The American Bankers Association Retail Bank Credit Report
gives annual statistics covering net loss rates on consumer lending
at commercial banks and also the proportion of losses associated
with bankruptcy, a part of which would very likely have been
written off without bankruptcy. For the years 1978 to 1986, banks
charged off about 0.5% of their closed-end installment loans in a
year. Losses identified with the borrowers' bankruptcies are in
the range of 15% to 24% of that 0.5% charge-off. Bankruptcy
losses were, therefore, about 0.1% to 0.2% of their outstanding
Although banks lose more on credit cards, and a higher propor-
tion of those losses are due to bankruptcy, still the credit card
losses due to bankruptcy appear in recent years to range from
0.32% to 0.51%. The credit card business is one of the most prof-
itable lines of business in banking and has been a major growth
segment in the lending business.
It was recently reported that during the first quarter of 1990,
more than $6 billion of bonds backed by credit card debt were
issued, most witR the highest (Triple-A) ratings. 8 Although the
authors speak of the pathology"" of the consumer credit market
and its fragility,"9 from these recent reports one would infer that
the large institutional buyers of such bonds judge this entirely
unsecured credit card consumer debt to be healthier and sturdier
than real estate backed bonds and most business firms' own
186. 1981-82 Hearings, supra note 165, at 565-66; see also infra Table 15.
187. Luckett, supra note 67. Luckett adds that the Federal Reserve's Annual Functional
Cost Analysis has higher credit loss ratios "but are still reasonably close and show the
same patterns year to year." Id. at 598.
188. Wall St. J., Mar. 26, 1990, at C1, col. 4. The reporter also related that one large
bank's own bonds were trading at a discounted price with a yield of 12.35%. "Yet the bank
successfully and easily sold $340 million of credit-card securities . . . yielding 9.37%." Id.
at C21, col. 1.
189. SWW, supra note 1, at 3.
190. Id. at 8.
191. A recent $1.25 billion bond issue backed by credit card receivables yielded 9.46%,
or only about 88 basis points more than U.S. Treasury securities. See N.Y. Times, June 22,
1990, at D14.
HeinOnline -- 43 Rutgers L. Rev. 231 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
Losses on Consumer Credit Extended by Banks
and the Proportionof Such Losses
Due to Bankruptcy, 1978-86
Losses net of Percent of Losses due to
Year recoveries as a losses due to bankruptcy as
percent of bankruptcy' a percent of
Closed-end installment loans
1978 .... .38 15.1 .06
......... . .44 16.4 .07
1980 .... . ......
... .63 19.9 .13
1981 .... .50 23.1 .12
1982 .... .40 23.6 .09
1983 .... .36 21.8 .08
1984 .... .28 21.9 .06
1985 .39 23.0 .09
1986 .... .. . . . .
.47 22.7 .11
198 1 ..............
* Source: Luckett, Personal Bankruptcies, 74 Fed. Res. Bull. 591, 598 table 2 (1988).
1. The ABA reports data for five asset-size categories of banks, but no overall average figure for all
banks. For this table, the average for all banks was caluclated by weighting the ABA's figure for each
size group by the proportion of total consumer installment credit (or total bank-card credit, as appro-
priate) held by the given group.
2. Calculated for this table by multiplying column 1 by column 2.
n.a. - not available.
HeinOnline -- 43 Rutgers L. Rev. 232 1990-1991
1990] BOOK REVIEW
During the years 1978 through 1986, "loss rates . . . have not
been the prime determinant of variation in net earnings on
closed-end installment loans. Net earnings have fluctuated con-
siderably more than have credit losses for this type of lending."1"
The cost of money to consumer lenders, their interest expense, is
a far larger factor in profitability than the loans written off. 1"
Luckett concludes that, with losses at 0.1% to 0.5% of con-
sumer credit receivables which constitute less than a fifth of total
bank lending, even more personal bankruptcies will make little
difference. The incremental loss rate has an almost negligible im-
pact on profitability. "[C]reditors may well have little incentive to
change lending standards in view of the relatively small impact of
bankruptcies on earnings and the uncertain outcome of a shift in
strategy."1 9 '
C. Comparing the Consumer Creditors
The authors claim that "[c]reditors in consumer bankruptcy
have been analyzed so seldom that we were initially unsure how
to sort them for comparison. ' ' 195 They need not have hesitated for
an answer. Quite apart from the comparisons in Table 16 below,
GAO broke down creditors by average total amounts into eleven
categories for unsecured credit and six categories for secured
credit and also into priority credit, business- and nonbusiness-re-
lated credit." '
CRC treats mortgage debt separately and has for non-mortgage
debt nine categories for secured credit and eight for unsecured
credit. These are also shown together.' 7 Nine Sts. and NJ Debt-
ors each have a breakdown for the five common types of lenders
engaged in the consumer credit business.'
192. Luckett, supra note 67, at 598.
193. For the first nine months of 1989, Beneficial Corporation's "lending spread," the
difference between interest expense (9.4%) and gross yield (18.5%), was about 9.1%. De-
linquent loans were 0.72% of the total portfolio. Barron's, Jan. 22, 1990, at 41, col. 1.
194. Luckett, supra note 67, at 598-99.
195. SWW, supra note 1, at 293.
196. GAO, supra note 1, 51.
197. 1I CRC, supra note 1, at 28-29 (exhibit 2-4 shows unsecured non-mortgage debt
and exhibit 2-5 presents secured non-mortgage debt).
198. NJ Debtors, supra note 1, at 564 table 25, shows the frequencies of all unsecured
creditors in 12 categories; Nine Sts., supra note 1, at 295-300 tables O-W, showed nine
such groups. The SWW study uses a similar listing for all creditors. See SWW, supra note
1, at 275 table 14.1.
HeinOnline -- 43 Rutgers L. Rev. 233 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
Table 16 compares the findings in average dollar amounts of
the most common consumer creditors and the percentage of the
average total of unsecured debt for all debtors.
J -- t C9
C'i Co Co
Or. 0 0
: Z5 4
_;a 4 r: 0
HeinOnline -- 43 Rutgers L. Rev. 234 1990-1991
1990] BOOK REVIEW 235
D. Creditor Legal Action
The authors relate that "[c]reditors had brought remarkably
few formal debt collection actions against these debtors prior to
bankruptcy."' " Only about one-third had been sued and less
than 10% had their wages or other property attached. This anom-
alous finding probably results from their site selection since both
Pennsylvania and Texas prohibit wage garnishment. 00
The authors report that they tested "whether having had a re-
possession or a property seizure was a predictor of the decision to
enter Chapter 7 versus Chapter 13."'101 They "found no significant
association with choice of Chapter ... ",202 They did not, how-
ever, similarly test wage garnishment, evidently the most impor-
tant creditors' remedy in relation to bankruptcy. Such a test
probably would have been of little significance since only one of
their three site states allowed that remedy and they would have
few such events. Table 17 compares the six recent studies.
Lawsuits Against Debtors
Lawsuits Attachments, and
Against Debtors Wage Garnishments
CT 46% 17%1
NJ Debtors 45% 20%
CRC - 33%2
Nine Sts. 40% -
GAO 44% 36% 3
SWW 33% 10%
1. CT, supra note 1, at 11.
2. Four percent of the CRC respondents (those debtors who were interviewed) identified wages
actually garnished as one of the two most important factors leading to bankruptcy. II CRC, supra
note 1, at 37 exhibit 2-9.
3. GAO, supra note 1, at 22 (allowing for no overlap in the 27% and 9% figures).
E. Income to Debt Ratios; Mortgages
The SWW study sometimes uses gross debt (including mort-
199. SWW, supra note 1, at 305.
200. NJ Debtors, supra note 1, at 586 table 48, shows wage garnishment in 11% of the
cases. CT had reported a 17% frequency of wage garnishments. CT, supra note 1, at 11.
201. SWW, supra note 1, at 269 n.37.
HeinOnline -- 43 Rutgers L. Rev. 235 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
gages) to income ratios. It is preferable and follows past practices
to treat mortgage debt separately.0 3 Including mortgage debt in
the bankruptcy-determinative ratio of debt to income entails ad-
ding some allowance for rent payments by the non-homeowner
debtors. Over the past several years the line has blurred between
long-term mortgage debt and the usually much shorter term con-
sumer debt.2" Even if the consumer debt is secured, the collateral
is personal property with nothing like the usually appreciating
lifetime value of residential real estate. Motor vehicles are now
financed for forty-eight to sixty months and mobile homes, even
longer. But the purchase money real estate mortgage is for most
families by far the largest credit transaction for the longest term,
often thirty years or more. Also, unlike personal property, resi-
dential real estate has appreciated in value during most of the
Over at least the past forty years, there has always been an ac-
tive market in shorter-term second mortgage lending. Consumer
creditors have always wanted a lien on the borrower's home as
security if possible. " 5 Calling these debts "home equity loans," as
we now often do, does not change the character of these second
mortgage loans. In regard to liens on real estate, one of the au-
thors' sillier commepts is submerged in a footnote: "Prior to the
development of these data, creditors might not have realized just
how valuable a security interest in the debtor's home is to protect
against the risks of bankruptcy. 20
F. Their Material on Consumer Credit Concludes with Two
More Fantastic Statements
We have taken a first step in introducing facts about creditors
and their lending practices into the policy debate. We began
203. See, e.g., CRC, supra note 1.
204. The Federal Reserve Board defines consumer credit separately from mortgage
credit. Consumer credit means all short- and intermediate-term credit to finance the
purchase of goods and services for personal consumption.
205. Even in the 1967 Hearings, there was concern over the excessive refinancing of
homes to finance the purchase of non-housing items. See 1967 Hearings, supra note 98, at
206. SWW, supra note 1, at 326 n.9. A recent article on Beneficial Corporation reported
that second mortgage loans comprised two-thirds of the company's loan portfolio. Bar-
ron's, Jan. 22, 1990, at 40, col. 3.
CRC had reported in 1982 that "[slecured credit such as second mortgage loans has
been substituted for unsecured cash loans." II CRC, supra note I, at 140.
HeinOnline -- 43 Rutgers L. Rev. 236 1990-1991
1990] BOOK REVIEW 237
• . . by noting that the creditors have been a faceless chorus in
bankruptcy debates focused almost entirely on debtors. 0
The first assertion should be considered in light of the follow-
ing list of some of the governmental studies of the consumer
credit industry and its firms, their practices, and the applicable
law: (1) Four volumes of 1967 Hearings on the "Consumer Credit
Industry" before the Senate Committee on the Judiciary; (2) the
Consumer Credit Protection Act of 1970 with Hearings before the
House and Senate Committees going back to 1967; (3) The Con-
sumer Credit Protection Act Amendments of 1977, with Hearings
before the House Committee on Banking, Finance and Urban Af-
fairs in five parts; (4) the Report of the National Commission on
Consumer Finance in 1972 on Consumer Credit in the United
States; and (5) the 1978 and 1980 Federal Trade Commission Re-
ports on Credit Practices based on staff analyses, studies, and ex-
tensive public hearings.
There is nothing faceless about these creditors. They have ap-
peared and testified at nearly all the legislative hearings on per-
sonal bankruptcy in the Congress and in the state legislatures on
opt-out legislation. They have commissioned studies and used va-
rious findings and claims to promote changes in the bankruptcy
law. In the 1981 House Hearings'"' eight groups of witnesses
presented statements and testified on behalf of various groups of
consumer creditors; all came with detailed statements and some
with commissioned studies. In the 1981-82 Senate Hearings,20 9
there were ten groups of witnesses for the consumer credit indus-
try. The CRC study was first publicly presented, along with
smaller studies of how personal bankruptcy affected various con-
G. Consumer Credit as Commercial Credit
The authors claim to explain the differences in the practices of
creditors in the consumer and commercial settings.2 This infor-
mation is not from their findings. They provide no basis for their
statements despite much published research on credit to small
207. SWW, supra note 1, at 322.
208. 1981-82 Hearings, supra note 165.
209. Bankruptcy Reform Act, 1978: Hearings Before the Subcomm. on Courts of the
Senate Comm. on the Judiciary,97th Cong., 1st Sess. (1981).
210. SWW, supra note 1, at 285, 287.
HeinOnline -- 43 Rutgers L. Rev. 237 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
Although they "cannot separate the entrepreneurs' debt into its
commercial and consumer components," they claim that their
data "show that a substantial amount of lending to individuals is
commercial lending."'2 How do they proceed from this first bar-
rier to the second proposition, which is the basis for some wide-
reaching conclusions? They assert that the differences in the av-
erage amounts of debts enable them "to draw inferences about
which creditors have extended commercial credit to the entrepre-
neurs."'1 s The debts of those to whom they refer as the "ever self-
employed" are larger than the loans to those who reported no
self-employment. The authors assume that some part of the
larger average debt is the commercial loans. Those "ever self-em-
ployed" are 20% of their sample. They suppose that debt totaling
much more than 20% of the total is commercial. From the infor-
mation provided by the SWW study, we cannot know whether
their inference is warranted or supported at all. It is an unproven
assumption on the basis of which the authors conclude that most
small commercial credit is unsecured. They also assert that many
academics are wrong in assuming that incorporation insulates
small business owners from personal liability. This is a strawman
argument which is discussed and dismissed later.'
XXI. WHAT THEY CLAIM TO HAVE LEARNED FROM THEIR DATA
1. "Bankrupt debtors are a cross section of America."218 The
authors intend to put us on, to have a joke with us. Although
there is great variation, and there are some sports, as there will be
in any group of persons that large, every study cited concludes
that the chapter 7 debtors are the near-poor: not impoverished
but far from the middle class.'
2. There is a very small group who abuse the system.'" There is
no reason to accept or reject their definition of abusers as (1)
211. See, e.g., SMALL BUSINESS FINANCE: SOURCES OF FINANCING FOR SMALL BUSINESS (P.
Horvitz & R. Pettit 1984) (two volumes with bibliographies).
212. SWW, supra note 1, at 282.
214. See infra notes 221-22 and accompanying text.
215. SWW, supra note 1, at 328.
216. The quoted statement must be rhetorical. SWW details the incomes, debts, and
assets of their sample of debtors, which is similar to those in the other empirical studies.
Id. at 64-65.
217. SWW, supra note 1, at 329.
HeinOnline -- 43 Rutgers L. Rev. 238 1990-1991
1990] BOOK REVIEW
those of their chapter 7 sample who could pay taxes and their
creditors in full and whose remaining income would be at or
above the BLS intermediate budget and (2) those who could pay
and have 80% or more of their take-home pay left. These two
groups make up about 5% of their chapter 7 debtors." ' Again,
although not mentioned, they use 1980 incomes which are com-
pared to 1981 BLS benchmarks, thereby substantially distorting
3. Women are worse off financially than men and earn less.
These facts are reflected in bankruptcy. Earlier studies had
4. "One discovery in these data is a group of formerly self-em-
ployed debtors."' About 10% were in business at filing and an-
other 10% were formerly self-employed. This newly found dis-
tinction, according to the authors, should be part of the published
bankruptcy statistics. 0 Like so much of the book, this is not new
information and should not have been called a discovery.
5. They say that "[many academics have dismissed the finan-
cial problems of small business owners, suggesting that they are
protected by corporate limited liability so that personal discharge
is irrelevant in coping with business debt. Our data thoroughly
dispose of that complacent assumption."" No citations are given
because these are strawmen statements. Academics seem well
aware that much small business credit is guaranteed by the own-
ers and their spouses, who are often sureties for the corporate
218. See id. at 211 table 12.3.
219. Id. at 330.
220. Nine Sts. had recommended that another category be created for the business-re-
lated personal bankruptcy filings. "[T]he known cases strongly suggest that lumping all
personal bankruptcy filings together leads to distortions and, for better description, the
business-related filings should be treated separately." Nine Sts., supra note 1, at 306.
221. SWW, supra note 1, at 330.
222. See Nimmer, Negotiated Bankruptcy ReorganizationPlans:Absolute Priorityand
New Value Contributions,36 EMoRy L.J. 1009, 1013 (1987) ("Even if there is a corporate
format, the individual owner often has most of her wealth tied up in or encumbered by the
See also Herbert, Consumer Chapter 11 Proceedings:Abuse or Alternative, 91 CoM. L.J.
Even if the [business] is incorporated, the informality common to very small
businesses virtually ensures that corporate and individual funds are commingled
and corporate debts individually guaranteed. . . . (It is generally true that the
joint filings of coordinate Chapter 11 proceedings by a small corporation and its
principal is . . . unremarkable, because the principal's major obligations (e.g.,
HeinOnline -- 43 Rutgers L. Rev. 239 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
6. Burgeoning consumer credit and increased volatility, they
say, are the two primary systemic factors that caused the "star-
tling increases" 28 in consumer bankruptcy in the 1980's. But
these increases are less than in other recent periods 2 ' and are
only about one-sixth more than the increases in business bank-
ruptcies, which suggests that consumer credit is not the cause of
the increase in personal bankruptcy filings.2"'
They "suggest a close connection between rising debt/income
ratios and rising numbers of bankruptcies. ' ' 22 6 This relationship
has been discussed in nearly all the literature on the causes and
correlates of personal bankruptcy filings for more than twenty
years. The 1967 Hearings addressed the issue in some detail and
with great concern 2 as did Brookings in 1971.228 Their suggestion
is not new or different and, most importantly, they offer no proof.
There is no check of their suggested correlation over time, despite
the existence of such data from some forty years ago, 22 ' nor is
there even any present comparison. Table 128' compares the ra-
those ubiquitous bank guarantees) arose out of her business activities and are
inextricably linked with the obligations of the corporation. Id. at 235.
223. The "startling increases in consumer bankruptcy filings in the 1980's," of which the
authors speak, SWW, supra note 1, at 12, are that bankruptcies per thousand persons rose
from 1.25 in 1980 to 2.03 im 1987. Luckett, supra note 67, at 595 table 1.
224. The 1967 Hearings used data from 1945 to 1965. Those showed that nonmortgage
consumer credit increased by 1500% while bankruptcy filings increased by 1400%. 1967
Hearings, supra note 98, at 48. This is a larger increase in bankruptcy filings, proportion-
ately, than over the ten years from 1979.
225. Over the decade from 1979 to 1988, there was a total increase of about 267% in
non-business bankruptcy filings, from 196,976 to 526,066. That this increase is not due to
consumer credit may be inferred from the increase in business filings during the same
time. From 1979 to 1988, business filings increased from 29,500 to 68,501, about 232%.
Reports of the Proceedings of the Judicial Conference of the U.S., Annual Report of the
Director of the Administrative Office of the U.S. Courts 146 (1979), id. at 31 (1988).
226. SWW, supra note 1, at 331.
227. Over the period from 1949 to 1966, the average debt service as a percentage of
family income about doubled. 1967 Hearings,supra note 98, at 102, 48.
228. When the Brookings Institution had completed its empirical study of the bank-
ruptcy system, the question of the causes of personal bankruptcy remained. One member
of the Brookings group apparently could not discern any significant differences between
persons who filed in bankruptcy and others, in similar financial situations, who did not
file. He posited, "a propensity to file in bankruptcy," his admission that he did not know
what caused bankruptcy filing. Brookings rejected his draft and the chapter was rewritten
to explain personal bankruptcy filing by examining income to debt ratios. A range of this
ratio may be a fact about the debtors who filed, but it does not explain why their cohort
group did not file.
229. Future Earnings Hearings, supra note 33, at 188-189 table II, shows personal
bankruptcy filings and debt-income ratios by year from 1950 through 1980.
230. Supra Part V.
HeinOnline -- 43 Rutgers L. Rev. 240 1990-1991
1990] BOOK REVIEW
tios of nonmortgage debt to pretax income in previously pub-
lished studies. The other three recent studies cluster very closely.
The apparently anomalous findings of SWW may be due to their
combining chapter 7's and chapter 13's for this purpose.
7. The second of their claimed causes of increased bankruptcy
is "the increased volatility and instability of economic life, an ac-
celerating change to which" everyone, it seems "has failed to ad-
just. '281 They state that since about 1971, "the American econ-
omy has become more volatile and unstable . . . ."' What do
these propositions mean? How can we know whether they are
true? The authors give us no measure and no history of volatility
8. Many debtors and many creditors are irresponsible in mat-
ters of borrowing and lending. The creditors lend too much and
the debtors borrow too much. Both groups, but especially the
lenders, should know, according to the writers, that repayment is
unlikely. The consumer creditors' net writeoffs strongly imply
that they are wrong.
9. They contend that they have proved that the legislation to
make chapter 13 more attractive has had little of its intended ef-
fect, and has generated some undesirable results. The authors
have not provided enough information, in particular, nothing on
payments to creditors, upon which to base a considered judgment.
10. Bankruptcy "provides a safety valve to avoid the formation
of a deeply angered debtor class. 23 How do they know that?
There are no data in their book to support the existence of this
potentially far-reaching threat to the Republic. This conclusion is
distressing. It suggests to me that, despite their protestations
("Our study is devoted to facts.. . .,),234 they are engaged in high
level populist theorizing that bears no relation to their research.
XXII. OMISSIONS IN THE SWW BOOK
"[Wie break new ground by devoting considerable attention to
the position of creditors in bankruptcy." ' I return to this claim
not merely because it is false, but also because the writers do not
adequately consider matters about creditors' behavior which are
231. SWW, supra note 1, at 332.
233. Id. at 340.
234. Id. at 341.
235. Id. at 6.
HeinOnline -- 43 Rutgers L. Rev. 241 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
important and matters which are important to creditors. Consid-
ering their claim to have focused on the consumer creditors for
the first time, etc., they hardly discuss what most others have
taken as matters of interest and concern.
(1) Most scheduled creditors do not file proofs of claim. " " The
authors mention this in passing2" but provide no information and
no discussion of the significance of so many scheduled creditors'
failure to file proofs of claim. This is vital information about
(2) The SWW study refers also to creditors' objections to dis-
charge"' and their efforts to have their debts excepted from dis-
charge.M But they provide no data on the frequency, the results,
or the effects of such proceedings. These issues have been of great
concern to creditors and to Congress.
(3) They estimate that about 19% of their wage-earning debt-
ors in chapter 7 reaffirmed one or more debts and they found
considerable evidence of informal reaffirmations, which were at
least as prevalent as formal ones.2 0 Most of the reaffirmations
appeared to them to be of secured debt.
Reaffirmations are very important to creditors as the authors
point out. Yet they do not tell us how much debt was reaffirmed
and which types of,creditors benefited from reaffirmations. The
issues in reaffirmation were a major concern in the drafting of the
Bankruptcy Code and its amendments.
(4) Exemptions, they say, "are often the key issue" at Code sec-
tion 341 initial creditors' meetings "because many debtors claim
all their assets as exempt, leaving nothing for liquidation and dis-
tribution among their creditors.''
236. "Of the 17 or 18 scheduled creditors to whom notice is sent, only three or four...
file . . . . Hence the creditors who do file proofs of claim sometimes get surprisingly large
dividends." Shuchman, Little Bankruptcies in New England, 56 B.U.L. REv. 685, 702
Brookings found that proofs of claim were filed only for a small percentage of the sched-
uled debts. Brookings, supra note 1, at 88-89.
GAO concluded "lpjroofs of claim are not always filed by secured creditors." GAO,
supra note 1,at 57-58. There were no filings for about 38% of the scheduled total amounts
in GAO's sample of chapter 13's.
237. SWW, supra note 1, at 81 n.20.
238. The authors say they found no "creditor objections to a debtor's discharge based
on the allegation of fraud." Id. at 265 n.11.
239. Id. at 32, 265 n.11, 277-78.
240. Id. at 32, 319.
241. Id. at 27.
HeinOnline -- 43 Rutgers L. Rev. 242 1990-1991
1990] BOOK REVIEW
(5) They tell us nothing about distribution to unsecured credi-
tors. How much was paid out to those creditors in chapter 7242 or
13 we do not know.
(6) Their data on the use of exemptions, they tell us, "are not
entirely clear, but it is fair to say that debtors in our sample often
escaped bankruptcy with somewhat more of their property than
the law allows."2 4"
They do not tell us how this was accomplished, which property
was claimed as exempt and, more importantly for the creditors,
for what amounts. All we are told is that in a set of 938 chapter
7's and 531 chapter 13's, an average of $14,498 in exempt prop-
erty was listed.2 4 '
(7) In order to claim all assets as exempt in 1981, many debt-
ors, especially those in Pennsylvania, with then sharply-limited
state property exemptions, would have had to use the spillover
provisions of the federal exemption law.2 45 Although there is some
indirect reference to spillover, 2 " without using that word, they
provide no information about its frequency of use or its useful-
ness by type of property and dollar amount.
This is not the happy carping of a critic that the authors should
have written a different book. All of the information in the pre-
ceding list was reorded in the bankruptcy files which were ex-
amined for their undertaking,247 much of it on the same official
forms. So I do not understand why they say that "it will take
another study to compare directly how much repayment goes on
in Chapter 7 and Chapter 13 .... "I'll
This book contains so much exaggeration, so many question-
able ploys, and so many incorrect statements, that it would be
well to check the accuracy of their raw data, as old as it is. But
the authors arranged matters so that they could not provide ac-
cess to the computer printouts by case, with the corresponding
242. Dr. Sullivan's letter advises that most of their sampled chapter 7 cases were
stamped "No Asset." Letter, supra note 44.
243. SWW, supra note 1, at 30.
244. Id. at 260 table 13.7.
245. The common practice of lawyers in the NJ Debtors study was to apply any unused
balance from the Code's "spillover" provision, 11 U.S.C. § 522(d)(5) (1988), to whatever
property had been claimed as exempt generally or by a specific provision of § 522 of the
Bankruptcy Code. NJ Debtors, supra note 1, at 542-43.
246. SWW, supra note 1, at 29.
247. Perhaps only the payments to creditors in their sample of chapter 13's would re-
quire calculation (or inquiry with the chapter 13 trustee).
248. SWW, supra note 1, at 339.
HeinOnline -- 43 Rutgers L. Rev. 243 1990-1991
RUTGERS LAW REVIEW [Vol. 43:185
bankruptcy court file numbers," ' thus preventing any indepen-
dent check of the raw data in the files from which they took their
This book is incomplete in other ways. The authors should
have data on such matters as (1) how much was paid to the credi-
tors in the confirmed chapter 13 plans, (2) the number and size of
reaffirmations, (3) the frequency of requests to have debts ex-
cepted from discharge with the results tabulated by type, credi-
tor, and average amounts, (4) the use of exemptions under state
laws and federal law and (5) the use of the federal exemption law
If the National Science Foundation had but known, it could
have saved the taxpayers all that money by commissioning an in-
expensive review study: What general statements about personal
bankruptcy may be tentatively accepted as true based on recently
published studies? The NSF could also have called for a meta-
analysis, combining the four similar studies for a quantitative
synthesis of the available data in those studies. This, too, would
have been much less costly than the authors' replication.
249. "Although the files we used are public documents, the petitioners ... had become,
for the social scientist, human subjects, and we used normal safeguards for the protection
of human subjects. These safeguards included the use of identifying numbers rather than
names on all the date tapes prepared, and careful storage, under lock, of any data with
identifying information." Id. at 350.
The exempt categories in The Department of Health and Human Services' "Policy for
Protection of Human Subjects" includes "[r]esearch involving the collection or study of
existing data, documents, and records, . . . if these sources are publicly available .... "
45 C.F.R. § 46.101(b)(5) (1989). "Human Subject" is defined as a person about whom a
researcher obtains "data through intervention or interaction with the individual . I...
§ 46.102(f)(1). The authors had no contact with any debtors. Hence the debtors are not
human subjects for this purpose and these court records are in an exempt category.
250. A common instance of misconduct in science occurs when "there [is] no way to
verify whether or not [the] research was accurate." Woolf, Deception in Scientific Re-
search, 29 JURIMERICS J. 67, 83 table 5 n.4 (1988).
HeinOnline -- 43 Rutgers L. Rev. 244 1990-1991