A Fighting Chance

housemaids who lived at the economic margins and always would: Many scholars once assumed that all debtors were “a poverty-stricken, chronically unemployed segment of the lower class”—people on the economic margins. Teresa A. Sullivan, Elizabeth Warren, and Jay Lawrence Westbrook, As We Forgive Our Debtors (1989), 63; see also Philip Shuchman, “Social Science Research on Bankruptcy,” Rutgers Law Review 43 (1990): 185.

the kind of study that legal experts almost never did: We were not the first researchers to try to understand more about the actual individuals who file for bankruptcy. For example, David Stanley and Marjorie Girth of the Brookings Institution conducted a notable study in 1971 that provided crucial insight into the sources of financial stress for those going into bankruptcy, the demographics of debtors, and debtors’ postbankruptcy experiences. See David T. Stanley and Marjorie Girth, Bankruptcy: Problem, Process, Reform (1971).

building a giant mosaic, one tile at a time: The 1981 study relied on bankruptcy petition data. The petition data included information regarding debts and assets, income, and the petitioner’s business, but it contained very little demographic information. The study covered districts in Illinois, Pennsylvania, and Texas. We coded financial information and supplemented our data with interviews with bankruptcy judges and bankruptcy lawyers.

The 1991 study relied on one-page questionnaires in addition to financial data from court records, which were vetted by the University of Texas and the University of Pennsylvania for the protection of human subjects. These questionnaires were designed to provide demographic and employment information about debtors. We asked trustees to distribute these questionnaires to debtors in Section 341 meetings, which are mandatory for debtors, in the first half of 1991. Debtors were told that participation in the study was voluntary and anonymous. We received about 59,000 questionnaires, which we winnowed down—through random selection—to a final sample of 150 cases from each district. In addition to drawing from the districts in the 1981 study, we drew from select districts in California and Tennessee. After coding the questionnaires according to a set of common criteria, we analyzed the data with an eye toward statistical significance while accounting for the possibility of selection bias and other types of data distortions. We supplemented this data with other publicly available data, such as the 1990 Census, as well as data generated by bankruptcy judges and economists.

As in the 1991 study, the 1999 study relied on core financial data and questionnaires that were distributed during mandatory debtor meetings. The study had an expanded geographical scope, covering districts in California, Illinois, Kentucky, Ohio, Pennsylvania, Tennessee, Texas, and Wisconsin. Our final sample for the 1999 study amounted to 1,496 cases. Again, we took steps to account for the possibility of selection bias and other types of data distortions.

In all three studies, we adhered to strict confidentiality protocol to protect the identities of participants.

For more information on the methodology underlying these studies, see the Appendices to As We Forgive Our Debtors (1981 study) and Jay Lawrence Westbrook, Elizabeth Warren, and Teresa A. Sullivan, The Fragile Middle Class (2000) (1991 study) as well as the 1999 supplementary study documented in Melissa B. Jacoby, Teresa A. Sullivan, and Elizabeth Warren, “Rethinking the Debates over Health Care Financing: Evidence from the Bankruptcy Courts,” NYU Law Review 76 (2001): 375.