Chapter 11 of the Bankruptcy Code presumes that managers will remain in possession and control of a corporate debtor. This presents an obvious agency problem: these same managers may have gotten the company into trouble in the first place. The Bankruptcy Code thus includes checks and balances in the reorganization process, one of which is supposed to be an \u201cexaminer,\u201d a private individual appointed to investigate and report on the debtor\u2019s collapse.<\/p>\n
We study their use in practice. Extending prior research, we find that examiners are exceedingly rare, despite the fact that they should be \u201cmandatory\u201d in large cases ($5 million in debt), and are recommended in all, if \u201cin the interests of creditors.\u201d Using a hand-collected dataset (n=1225) of chapter 11 bankruptcies from 1991-2010, we find that they are sought in less than 9% of cases (104), and appointed in fewer than half of those (48, or 3.9% of the sample).<\/p>\n
We offer three observations about the under-use of examiners. First, regression modeling shows that the factors that predict when an examiner will be appointed have little to do with the agency problems that concerned Congress, such as fraud or mismanagement. Instead, the timing of an examiner request and case venue appear to be the most important factors in the rare cases where they appear. Second, governance in reorganization has changed significantly since Congress enacted chapter 11, yet agency problems persist. Rather than reorganize \u201cin place\u201d with operational management, troubled companies are increasingly captured by sophisticated distress investors (e.g., hedge funds), who use \u201cturnaround managers\u201d to enhance their recoveries. Examiners could tell us whether this change has net social costs or benefits \u2014 if they were used. Third, we offer preliminary evidence that examiners should be used more frequently, because the rare case with one is likely to be more \u201csuccessful\u201d in a variety of ways than a case without one.<\/p>\n