More Women in Bankruptcy1


Teresa Sullivan
Professor of Sociology and Vice-President and Dean of Graduate Studies
University of Texas at Austin
Elizabeth Warren
Leo Gottlieb Professor of Law
Harvard University

Prepared for the American Bankruptcy Institute
Web posted and Copyright © July 30, 1999, American Bankruptcy Institute.

Our recent report of the rising proportion of women among filers for non-business bankruptcy has stimulated three additional reports of data, which we incorporate in this updated report. Professor Pollak has extended his earlier time series of filing data to include the most recent (1997-1999) filings in Nebraska, and researchers at the Executive Office for the United States Trustees and for the credit industry have re-analyzed their recently reported studies (1996-97). These additional data points strengthen our earlier analysis and reinforce our earlier conclusion: Since the early 1980s, the number of women filing for bankruptcy alone has risen at a much faster rate than for men filing alone or for married couples.

This site recently headlined the re-analysis of the Ernst & Young/Visa/MasterCard study as "disputing" our findings. Reporting on his analysis of the credit industry data, Mr. Thomas Neubig reports 32% women single filers among Chapter 7 filings in 1997. We report 39.2% women single filers among Chapter 7 and Chapter 13 filers two years later. There are differences, but so far no "dispute." The credit industry contributes data from one point in time to the data

series that we presented. The credit industry data alone do not permit an inference about whether women are increasing or decreasing their representation in bankruptcy because a trend cannot be inferred from a single point in time. When we add their data to the reports of independent researchers, however, a clearer picture emerges about women in bankruptcy.

Two other new reports help shed additional light on the trend toward increasing proportions of single women filers. Professor Pollack follows all non-business cases over an extended time in a single district (Nebraska), reporting particularly sharp increases in the proportion of women filing alone in 1998 and 1999. He explains that from July 1997 through June 1999, the proportion of women filers "increased incrementally from 32.79% to 33.82% to 34.43% to 36.28% in the last six month period ending June 30, 1999." Gordon Bermant and Ed Flynn, analyzing a data base from the Executive Office for the United States Trustee, add another data point to confirm the trend. They examined cases filed in the second half of 1998 and the first quarter of 1999 and found that 34.6% of the cases were single-filing women.

A plot of the reported data points, combining those findings reported earlier from our own research, Professor Pollak's research, and Professors Culhane and White's research along with all three recent entries, Professor Pollak's newer data, Mr. Neubig's re-analysis of the credit industry data, and Mr. Bermant and Mr. Flynn's re-analysis of the U.S. Trustee data, illustrates this trend:


1. 1981 data: Collected from 1,547 non-business bankruptcy cases filed in Chapter 7 and Chapter 13 in ten federal districts. The protocols for that study and the data are reported in full in Sullivan, Warren & Westbrook, As We Forgive Our Debtors: Bankruptcy and Consumer Credit in America (Oxford University Press 1989), 149. We report there that the proportion of single women filers in 1981 was 17%. Ed Flynn has published AO data showing that the national proportion of joint filings in 1981 was 43%, considerably lower than our ten-district figure of 57%. Ed Flynn, "Bankruptcy by the Numbers," ABI (June 1992). As we have noted previously, Texas, one of the states in our 1981 ten-district sample, had a high proportion of joint filings. We re-analyzed our data assuming a joint filing proportion of 43% and redistributed the remaining 57% according to the male/female proportion identified in our sample. This gives an estimate of 22.1% single-filing females. We show both estimates in the entries for 1981.

2. 1987 data: Collected from 5,441 non-business bankruptcy cases filed in chapter 7 and Chapter 13 in Nebraska. Oliver B. Pollak, "Gender and Bankruptcy: An Empirical Analysis of Evolving Trends in Chapter 7 and Chapter 13 Bankruptcy Filings 1996-1997," 102 Commercial Law Journal 102, 333-38 (1998)(reporting 22.2% women single filers).

3. 1991 data: Collected from 2,650 cases filed in Chapter 7 and Chapter 13 in sixteen federal judicial districts. The protocols for that study and the data are reported fully in full in Sullivan, Warren & Westbrook,The Fragile Middle Class: Americans in Debt (Yale University Press 1999) (reporting 30% women single filers). (This book will be available in the fall from Yale University Press, but the chapter on research protocols is available on request from the authors.)

4. 1995 data: Collected from 1,047 Chapter 7 cases filed in seven judicial districts in 1995. The protocols for that study are reported fully in Marianne B. Culhane & Michaela M. White, "Preliminary Results of the Bankruptcy Reaffirmation Project" (Oct. 1998)(unpublished manuscript on file with authors) and in Marianne B. Culhane & Michaela M. White, "Taking the New Consumer Bankruptcy Model for a Test Drive: Means-Testing Real Chapter 7 Debtors," 7 American Bankruptcy Institute Law Review 27 (1999) (supplementary analysis by Culhane and White, July 26, 1999, reporting 34.5% women single filers).

5. 1996-97 data: See Pollak, note 2 above (reporting 32.4%).

6. 1997 data: Collected from "more than 2,100" Chapter 7 cases self-reported as "nationally representative" but protocols for site location, sampling and weighting are not published. Letter to Samuel J. Gerdano, dated July 19, 1999, by Thomas Neubig, Ernst & Young LLP (reporting 32%).

7. 1998-99 data: Collected from 1,452 non-business, no-asset Chapter 7 filings from 65 districts from July 1998 through March 1999. Unpublished data compiled by Gordon Bermant and Ed Flynn, Executive Office for United States Trustees, fax from Ed Flynn to Dr. Sullivan, dated July 29, 1999 (reporting 34.6%).

8. 1999 data: Collected from 1,521 bankruptcy cases filed in Chapter 7 and Chapter 13 in eight districts during the first quarter of 1999. Protocols are reported in Teresa Sullivan and Elizabeth Warren, Women in Bankruptcy, ABI Web, July 14, 1999. In addition, in a review of 11,780 Nebraska cases, Oliver Pollak reports that from July 1997 through June 1999, the proportion of women filers "increased incrementally from 32.79% to 33.82% to 34.43% to 36.28% in the last six month period ending June 30, 1999." Fax from Mr. Pollack to Professor Warren, dated July 16, 1999; data reported in ABI . Mr. Pollak counts all non-business cases in Chapter 7 or Chapter 13, not just samples, from all filing locations within the state. See note 2 supra. Both data points are reported on the graph.

The increase in proportion of women at the same time that the absolute number of filings is growing means that the number of single-filing women in the bankruptcy courts is rising even more sharply. If the various reported proportions from all sources (U.S. Trustee, Culhane & White, E&Y, Pollack and Sullivan, Warren & Westbrook), could be projected to the universe of filers, and using the higher estimate of women's filings from 1981 data, the overall number of women in bankruptcy from 1981 through 1999 would be charted in Figure 2.


Professor Pollak, who has collected data on all the non-business cases filed in a single state over a several year period, expresses his concern about the accelerating proportion of women filing alone for bankruptcy. He notes that "women are increasing their lead as the largest group filing bankruptcy in Nebraska at the ominous rate of about 1.328% per year over the last 2.5 years."2

For people deeply immersed in empirical research, it is always interesting to speculate on why there are slight differences among the various reports. One obvious difference, of course, is that most of the data are collected at different points in time. All three post-1997 reports, Professor Pollak, Mr. Flynn and Mr. Bermant, and Professors Sullivan and Warren show higher proportions of women filing alone than Mr. Neubig reports for 1997. Of course, even within the same time period there are some differences: the 1999 reports from our data and Professor Pollack's data differ by about 2.9 percentage points.

Some of the differences among the reports may be attributed to sampling error. A difference between 34% and 38%, for example, may or may not be statistically significant. If researchers drew another random sample from their original pools, they might have a slightly different number. The only researcher to escape this problem is Professor Pollak who counted every case in his district.

A third source of difference is that the data bases are drawn from different groups of filers. The Pollak and Sullivan & Warren samples include both Chapter 7 and Chapter 13 filers, while Culhane and White, E&Y and the U.S. Trustee samples include only Chapter 7 filers. In addition, each research effort began with a different mix of judicial districts which could have some effect on the reports. Another possibility is sampling differences among all four studies, both in the number of districts and in how the data were drawn within each district.

Finally, the differences in how gender information is collected could be at work. The Culhane & White, E&Y, U.S. Trustee and Pollak samples are based only on the information in the bankruptcy files, while the Sullivan and Warren data include debtor questionnaires. Because petitioners do not report their sex, each researcher using only petitions must classify the single filers as either "men" or "women" and record that information. In bankruptcy samples with Leslie, T.J., Chris, Dae Joon, Thien V., Jimmie, Lynn, H.L., Seno, Travell, Sukunya, Shannon, Kim, Dana, Luz and Marion (actual names from two of the samples reported here), classification by gender can be somewhat difficult. In their analysis, Culhane and White report that when they looked at petition information, they could not determine gender for 1.2% of their single filers. Even when he read the full files of the debtors with ambiguous names to find a single reference to "she" or "he" in the file, Ed Flynn reports that he could not classify 1% of the cases. Mr. Neubig does not report any debtors for whom he could not determine the sex. There is, of course, the compounding problem that even a researcher who had no question about the sex of the petitioner may be wrong. Cameron, Campbell, and Alex are women recently in the news who might be easily classified as male single filers if they had appeared in a bankruptcy sample.

Because we had questionnaire data, we were able to ask the debtors to indicate their sex on our questionnaire. To check the representativeness of our questionnaire data, we pulled a random sample of 300 petitions in two of the same districts and classified them by joint petition, single petition male and single petition female. Couples are easy, but the researcher making the classifications had a question about the sex of 12 single filers out of a total of 180 single file cases.

In his contribution to the data analysis, Mr. Neubig suggested taking the discussion about women debtors in a different direction. He argued that it would be better to discuss these women in terms of their marital status. We believe that is an important question, and we are currently examining the debtor's descriptions as married, divorced, separated or widowed. When we have analyzed our data, we will report those data and make them available, and we hope others will be able to do the same. We make only one cautionary note now: a single-filing debtor who describes herself as "married" is not necessarily part of an intact economic unit. As we noted in our 1989 book, As We Forgive Our Debtors, "some single-filing women may be married at the time of filing, but only because the bankruptcy court is faster than the divorce court" (p. 150). We described in some detail the economic advantages to a couples' filing jointly, if they are in fact a marital unit. Indeed, we explained that some proportion of the joint petitions are from debtors who are taking their last economic step together--dealing with overwhelming debts as they try to forge new economic lives. In the complex lives of the debtors, a picture taken at a single point in time is always incomplete. The filing status of women is legally and socially important and is worth studying in its own right, whether these women are married, single, divorced or widowed.

Mr. Neubig ends his letter with a comment about the proportion of women that would be affected by proposed House legislation regarding a means test. He reports that fewer of the single-filing women in his sample will be swept into the means test than will couples, presumably because the women have lower incomes and higher expenses. He might be right that the single-filing women in bankruptcy are not only growing in numbers, but that they are in worse financial shape than their joint-filing counterparts--suggesting that single-filing women are economically distinct from joint-filing women. That economic disadvantage, however, does not translate into exemption from the effects of proposed changes in the bankruptcy laws. Because the overwhelming majority of the provisions apply to all debtors, regardless of income, the impact of increased filing requirements or reduced automatic protection would have a disproportionate impact on those with the fewest resources to protect themselves through pre-planning, extensive legal advice, and extensive record gathering. Mr. Neubig's data suggest the possibility that increases in debts treated as non-dischargeable, increases in the amount of repayments required on secured debt, increases in the costs of filing for bankruptcy, increased reporting burdens and literally dozens of other provisions that are not income-tested would fall hardest on women filing alone.

The credit industry data not only add evidence of a trend toward increasing number of single filing women, but they also add a strong indication that the women filing alone are in relatively worse financial shape than their married counterparts by the time they file for bankruptcy. If this can be confirmed, it would suggest that women single filers are struggling to stay out of bankruptcy as long as they can and filing only after they are in more desperate shape than their married counterparts. This is another issue we hope to explore as we continue the analysis of our data and one we hope others will examine as well.

We are very encouraged to see more people re-analyze their data to examine the issues about women in bankruptcy. We are also heartened by attempts to understand the bankruptcy process through empirical research, but we must raise a cautionary note. The GAO recently reported that it requested the reported data from Mr. Neubig (E&Y) and Professors Culhane and White in order to evaluate their differing reports on the repayment capacity of Chapter 7 debtors. According to the GAO, Professors Culhane and White made their data readily available, while E&Y refused, claiming that their data are proprietary to Visa/MasterCard, the sponsors that paid for it. We welcome a continuing dialog over women in bankruptcy and we note that the E&Y data reported here are generally consistent with independent data. Regardless of the seeming reliability of any particular report, however, we urge that public policy debates rely on independent data, not on proprietary data that are disclosed only at the discretion of those who have an financial interest in the outcome.


1. For a paper entitled "The Price of Equality? Women in Bankruptcy," we are examining empirical data about who has filed for bankruptcy in 1999 and their reasons for filing bankruptcy. ABI Director Sam Gerdano asked us if we would write a very brief summary of our initial findings. We agreed, as we emphasized that our early report would be confined to only a small corner of the newest data, and that we are continuing with the full analysis. Since we published that report, three additional researchers have published comparable data about the proportion of women in bankruptcy. Here we update our earlier report, Women in Bankruptcy, July 13, 1999. [Return to Text]

2. See Figure 2, Note 8 above. [Return to Text]