Less stigma or more financial distress: an empirical analysis of the extraordinary increase in bankruptcy filings.

• Author: Sullivan, Teresa A.

INTRODUCTION A. Prelude: Three Studies I. THE RISING BURDEN OF DEBT A. Income: Still Low B. Assets: Increasing as Home Values Increase C. Debts: Rising Fast D. Debt-to-Income Ratios: Confirming the Bad News II. THE DECLINING STIGMA HYPOTHESIS III. THE INCREASING STIGMA HYPOTHESIS IV. ALTERNATIVE EXPLANATIONS CONCLUSION APPENDIX INTRODUCTION

The passage of the 2005 amendments to the Bankruptcy Code represents the turning of a page in the long history of personal bankruptcy in the United States. (1) At the very moment that European countries are liberalizing their treatment of individual debtors, (2) the U.S. Congress has embraced changes intended to make bankruptcy difficult or impossible for many financially troubled Americans. The primary justification for this wholesale revision of the accessibility of the consumer bankruptcy system has been the repeated claim that the extraordinary increase in bankruptcy filings is the consequence of declining stigma. In effect, the argument is that a growing moral slackness causes people who can repay their debts to seek the too-easy protection of bankruptcy. This Article reports the third of three comparable empirical observations of individual bankruptcy spread over twenty years. (3) It establishes a baseline against which the effects of the new amendments can be examined. The data we present are not consistent with the claim that declining bankruptcy stigma has fueled an increase in bankruptcy filings. Instead, the data are far more consistent with the hypothesis that increased filings result from increased financial distress, and they hint that, despite loud claims to the contrary, the stigma of bankruptcy may actually be increasing.

Historically, official declarations of bankruptcy incorporated overt stigma. In European countries in the sixteenth and seventeenth centuries, a person declaring bankruptcy was required to engage in humiliating public behavior. The law in Padua is illustrative. The bankrupt was required to appear naked or nearly naked in the "vast Paduan Palace of Justice" and to slap his buttocks three times against "The Rock of Shame" while loudly proclaiming, "I DECLARE BANKRUPTCY." (4) It appears the reason for this requirement was a concern that bankruptcy might otherwise involve too little stigma. The traditional notion in many societies is that bankruptcy is, and should be, shameful and that stigma plays a critical deterrence role. (5)

Whether stigma has played an adequate role in deterring bankruptcy in the United States has been called into question. Over the past two decades, families declaring themselves busted, unable to make it to the next payday, have tumbled into bankruptcy in record numbers. In 1981, there were about 3.6 non-business bankruptcy filings for every thousand households in the United States, (6) for a total of 315,832. If the filing rate per household prevalent in 1981 had remained steady, the number of bankruptcy filings in 2004, the last full year before the bankruptcy laws were changed, would have been roughly 429,000. (7) In fact, by 2004 the rate of filings had surged to fourteen per thousand households, for a total of 1,563,145 families in bankruptcy--a new bankruptcy case every twenty seconds. (8)

As they pressed for a change in bankruptcy laws, credit card issuers asserted that declining stigma caused the extraordinary run up in bankruptcy filings. (9) Their ideas were quickly echoed in the popular media (10) and were enthusiastically taken up by moral critics who loudly proclaimed that bankruptcy had lost its stigma. (11) Congressional leaders joined in the chorus, adding a rare bipartisan voice as they denounced their own constituents as morally deficient. Democratic Senator Patti Murray (D-Wash.) claimed that the Senate should make it a priority "to recapture the stigma associated with a bankruptcy filing," (12) while Republican Senator Bill Frist (R-Tenn.) asserted that "[b]ankruptcy has become so common that it has lost the stigma it had even a short generation ago." (13) Federal Reserve Bank Chairman Alan Greenspan made the same assertion: "Personal bankruptcies are soaring because Americans have lost their sense of shame." (14)

Other economists took up the claim as well. (15) Unable to find a strong statistical correlation between bankruptcy and a handful of macroeconomic indicators, they attributed the unexplained rise in bankruptcy filings to the unmeasured concept that they conveniently labeled as a reduction in stigma. (16)

In this Article, we draw on data from three empirical studies of families filing for bankruptcy conducted in 1981, 1991, and 2001 to test the claim that bankruptcies have increased because stigma is on the decline. The data we offer are flatly inconsistent with the declining stigma hypothesis. While no single statistical analysis is dispositive, the data presented here provide strong reason to doubt that the stigma hypothesis correctly explains the rise in consumer bankruptcy filings. In fact, we suggest that these data support an alternative view that the stigma of bankruptcy has actually increased over the twenty-year period we have studied, and that bankruptcy filings may have risen despite increased shame about declaring bankruptcy.

Part I of this Article briefly sketches the demographic profile of American consumer bankruptcy, demonstrating again that bankruptcy is a middle class phenomenon, primarily employed by families near the middle of social and economic life in the United States. This Part also provides basic financial data about Americans who filed for bankruptcy at the beginning of the twenty-first century and compares them with those who sought bankruptcy protection in the last two decades of the twentieth century. It shows that the central characteristic of consumer bankruptcy over two decades has been increasing financial distress, marked by rising levels of debt. Part II uses these data to test the declining stigma hypothesis. Part III explores the possibility that stigma may be rising--a hypothesis consistent with the data presented here. Part IV identifies other data that may help explain a rise in consumer bankruptcy filings.

1. Prelude: Three Studies

   The Consumer Bankruptcy Study consists of three large studies of natural persons filing for bankruptcy in Chapter 7 or Chapter 13 in 1981, 1991, and 2001. These studies constitute a unique dataset covering nearly the entire period since the complete rewriting of the Bankruptcy Code in 1978. The methodological details we have followed in our three studies are set forth in detail in three books. (17) We coded information on income, assets, and debts from bankruptcy court records in each study. Those data are typically prepared with the help of a lawyer or paralegal, and they are always filed with the courts under penalty of perjury.

   In 1981, we supplemented our court-record data by interviewing bankruptcy judges and bankruptcy lawyers. In 1991 and 2001, we collected written questionnaires directly from the debtors. (18) In 2001, we also added telephone interviews with bankrupt debtors. The core financial information reported in this Article comes from the schedules filed with the court, and the written questionnaires used in 1991 and 2001 provide the primary basis for the demographic data reported in this Article, including information concerning gender, education, and occupation. (19) Telephone interview data are not included in this report. (20)

   Although these studies were not national in scope, each of them involved a number of federal districts and approximately 1550, 2650, and 2000 bankrupt families, respectively. In 1981, the financial information was collected from all ten federal districts spanning three states--Texas, Illinois, and Pennsylvania. For 1991 cases, we collected financial information from one district in each of those states, plus one district from our other two sample states for that year, California and Tennessee. The 2001 financial data were drawn from the same five districts as in 1991, except that we sampled the Northern District of Texas instead of the Western District. (21) The samples were drawn systematically in each year, reflecting the proportion of Chapter 7 and Chapter 13 cases for each district sampled. In 1981, respondents were sampled systematically from the court's docket for the year. Because debtors could not opt out from being included, the resulting bias was limited to the small likelihood of technical sampling error (or the chance that a different systematic draw of debtors would produce different results). In both 1991 and 2001, we approached debtors at their meeting with the trustee in bankruptcy (22) to obtain their cooperation with the study and administered a brief questionnaire. Subsequently, a systematic sample of these questionnaires was drawn to link to the public records. This procedure is subject to response bias, which is the chance that the respondents who spoke to us were systematically different from those who would not speak to us. However, a small study of the financial records of filers who did not agree to complete the questionnaire indicated minimal nonresponse bias in the financial data. (23)

   The 1981 data were collected for the whole year. After our analysis revealed no seasonal biases, we gathered data in 1991 and 2001 for the first half of the calendar year. This means that the 2001 questionnaires were collected well before the terrorist attacks of September 11. To make comparison across time periods easier, we have adjusted all the figures in text and tables to constant 2001 dollars. (24)

   1. THE RISING BURDEN OF DEBT

   When we began our first study in 1981, conventional wisdom held that bankrupts were day laborers and housekeepers, for the most part blue collar or lower. (25) To general surprise, including our own, we found that the debtors were solidly middle class. More than half went into bankruptcy owning their homes, and a large portion...