| It is well known that a large fraction of firms seeking to reorganize under the U.S. Bankruptcy Code will be liquidated instead, and that the liquidation decision will often be made by a bankruptcy judge. Little is known, however, about the characteristics of optimal judicial decision-making in this context and whether the behavior of actual judges is consistent with these characteristics. Despite the paucity of theory and evidence, it is widely suspected that judges are poor decision-makers who allow failing firms to linger under the protection of the court. Drawing on optimal stopping theory, this study develops several new models of optimal judicial decision-making. Each generates the same implications: the probability (or hazard rate) of shutdown should be hump-shaped over time and decreasing in the uncertainty (or volatility) surrounding a firm's going-concern value. These implications are tested against new data on decision-making by bankruptcy judges of the Northern District of Illinois, Eastern Division, a jurisdiction covering Chicago and outlying areas. The data include all corporate Chapter 11 filings during 1998; the vast majority involve small firms with fewer than 20 employees. The data support the theory, and do not support the conventional wisdom. Various empirical tests---tabular comparisons, simple discrete-choice models, and biostatistical duration models---are consistent with the implications of the theoretical models. These findings, and the additional finding that shutdowns occur rapidly (over 70% of shutdowns occur within the first five months of a case), are surprising and suggest that the behavior of bankruptcy judges (non-market actors) may resemble that of market actors. On the other hand, the findings raise some doubts about the benefits of attempting to reorganize small firms, most of which are thinly capitalized and easily replaced. Bankruptcy judges may act rationally, given the constraints of the legal system, but the system itself may be poorly designed. |
