Google+ Followers

Thursday, June 13, 2013

There is no Evidence that Unsecured Recoveries in Chapter 11 are Diminishing (Part 5: Changes in Subordination Methods and Use of First-Day Orders)

This is the 5th in a series of blog posts that analyze an argument, being presented to an American Bankruptcy Institute commission on reform of the Bankruptcy Code, that the recoveries of unsecured creditors are somehow declining due to a purported imbalance in favor of secured creditors.  The argument relies heavily on a law journal article published two years ago by a law student named Andrew A. Wood working under the tutelage of Professor Lynn LoPucki at UCLA.   So far, I have shown how the article contains, and thus the argument is based on, incredibly unreliable and error-riddled data; that the sample sets being contrasted are not properly comparable because they cover different lengths of time and different economic conditions; that the purported increase in the use of second lien debt in capital structures cannot be shown to have had any adverse effect on unsecured recoveries; and that valuation and intrinsic business merit have much more to do with low recoveries than inter-creditor balance does.   In this post, I will show that the Wood article's portrayal of declining unsecured recoveries results in large part from a sadly unsophisticated grasp of corporate finance and chapter 11 practice; more specifically, it omits from the tabulation of unsecured creditor recoveries all recoveries that occur outside a plan (for example under first-day orders), and it fails to comprehend that unsecured recoveries can vary based on structural subordination.

Likely Misclassification of “General Unsecured Creditors”

Wood has broken out “unsecured creditors” into multiple categories, as LoPucki did in his earlier article. Those categories are “Senior Unsecured”, “General Unsecured”, “Senior Subordinated” and “Junior Subordinated”.  Partly because there is no single classification regime that proponents have to follow in chapter 11, and partly because, I suspect, neither Wood nor LoPucki had much experience in negotiating chapter 11 plans, I fear that there may be significant mis-classifications in both studies, particularly between “Senior Unsecured” and “General Unsecured” which may have affected Wood’s calculations of average recoveries.   Wood states that he follows LoPucki’s earlier work in treating “all unsecured creditor classes to be general unsecureds unless words suggesting different priority, such as ‘subordinated,’, ‘junior’ etc.] were used” (Wood, p. 433).  That definition completely ignores, or obscures, the fact that, in cases involving multiple debt issuers within the same corporate family, absent substantive consolidation, unsecured creditors at different debtors may have different recoveries. In particular, where a holding company exists, debt issued there often has significantly lower recoveries than unsecured claims at one of its operating subsidiaries may have. Practitioners refer to the holding company’s debt as “structurally subordinated”.  However, Wood’s table of recoveries does not consistently distinguish and categorize the kind of unsecured claims that can be considered “general unsecured” – trade debt, senior unsecured notes and so on.  Sometimes he puts senior unsecured notes in the “general unsecured claims” bucket; sometimes he records recoveries only by one set of notes in a corporate structure with elements of structural subordination; and sometimes he ignores the notes’ recoveries altogether. 

His data for recoveries in the Lear Corporation case illustrate this inaccurate and inconsistent labeling.  His table shows a recovery of 100” for “Senior Unsecured” creditors and a recovery of “36-42%” for “General Unsecured”.  But Wood has misunderstood the classification of unsecureds in that case and as a result substantially mischaracterized the various constituencies’ recoveries, as I confirmed by examining the recovery table on page 15 of  the disclosure statement (docket 634).[1]  First, two groups of unsecured creditors recovered 100%: “Ongoing Operations Unsecured Claims” against the “Group A Debtors”, estimated to be $410 million; and “General Unsecured Claims” against the “Group B Debtors”, estimated to be $285 million. Second, “General Unsecured Claims” against the “Group A Debtors” estimated at $2.104 billion, were indeed estimated to receive 42%, but that class consisted almost entirely of (a) credit agreement deficiency claims totaling $737 million and (b) unsecured bonds totaling $1.29 billion.  Substantially all of the quintessential “general unsecured” creditors of Lear received a 100% payout.[2]  Worst, there was no class of unsecureds that received the “36%” payout included in Wood’s table; rather, that figure was erroneously copied by Wood from the liquidation analysis for general unsecured creditors of the Group B Debtors. 

An equally significant flaw with Wood’s table is its failure to recognize differences in unsecureds’ recoveries stemming from “structural subordination”.  I found numerous instances in Wood’s 42 cases where structural subordination of unsecured debt occurred but went unacknowledged, such that the structurally subordinated debt was classified as “senior unsecured” or “general unsecured”, even though there were significant disparities between recoveries of unsecured creditors at different levels of the corporate structure.  Cooper Standard, R.H. Donnelly, Simmons Bedding, Primus Telecommunications, NTK, Six Flags, and Charter Communications are all cases that I have discussed in earlier posts in which some class of structurally subordinate unsecured debt recovered less than more structurally senior unsecured creditors.  The occurrence of this phenomenon is so frequent that it ought to have been separately reported and analyzed. 

Further, notwithstanding the definition Wood recites, when I look at LoPucki’s study of 1991-1996 cases, I see a surprisingly large number of “100%” recoveries by “general unsecured creditors”, leading me to suspect that those are mostly trade and other claims that are not on account of long-term debt that happens to be unsecured.  I am skeptical that unsecured debt in those cases was consistently and accurately classified, given that it was not properly analyzed in the 2009-10 sample.  Instead, there may be substantial errors in, or inconsistencies between, the two studies in regard to how they classify unsubordinated unsecured debt when there are multiple debtors.  There is simply no way to tell if the studies are comparable, apples to apples, or even if similar claims are categorized consistently within the same study.  Thus, none of the bottom-line figures – whether they be “77%”, “53%” or 45” – are reliable.

I suspect that, were someone to have access to all the 1991-96 cases, one would see that a good bit of the purported difference in recoveries between the two eras can be explained just by this classification approach: the LoPucki study happened to cover an era with more instances of contractually subordinated unsecured debt, which, being junior, tended to have lower recoveries.  So, when his table segregated that kind of debt from the so-called “General Unsecured” category, arithmetically the latter showed a higher apparent recovery.  Whereas Wood, likely unaware that recent capital structures have comparatively little contractually subordinated debt but a meaningful amount of structurally subordinated debt,  has blindly applied the older taxonomy, unknowingly lumping a different kind of subordinated debt into the general unsecured category and thereby diluting that category’s recoveries.  But to prove that out would require one to examine all of the 1991-96 cases to see how each class of general unsecured was categorized, a task beyond the scope of this paper. 

At a minimum, the issue of failure to account for different modes of subordination consistently between the two studies illustrates how unreliable the two studies are, relative to how practitioners negotiate and understand the terms of chapter 11 plans. Further, stepping back from the errors, omissions and inconsistencies, a question is raised as to whether recoveries of unsecured creditors at a holding company that has no operating creditors and no secured debt are even relevant to the contention that secured creditors are usurping general unsecured recoveries.

Omission Of Other Payments To Unsecured Creditors. 

As practitioners know, there are numerous ways that unsecured creditors recover in chapter 11 cases that don’t show up in the recovery estimate tables in a disclosure statement.  There are payments of critical vendors; assumption of executory contracts and unexpired leases; and priorities of various priority claims for employees and retirees pursuant to first-day orders, to mention what are probably the largest ticket items.  There are many situations both under plans and in 363 sales where unsecured creditors “ride through” unimpaired economically, being assumed by the reorganized debtor or the acquirer.  Also, there are many situations in which an unsecured claim is insured, by a liability policy, for instance, or a letter of credit.  For example, in the first Journal Register case, where the debtor reported accounts payable of less than $19 million in its last 10-K before filing chapter 11, its newsprint supplier was owed approximately $2.7 million at the petition date, yet held a $3 million LC, according to the company’s critical vendor motion (docket 15).  Neither the LoPucki nor the Wood article includes any data or estimates about the recoveries unsecured creditors received through non-plan channels.

In many of the cases Wood sampled, I was struck by how small the “General Unsecured” class was under the plan.  Often it was a small fraction of the claims in the case, so small sometimes that it made no sense as a financial matter.  Most notably, in R.H. Donnelly, the General Unsecured class was only $19.5 million, according to the final disclosure statement (docket 463).  For a company with about $10 billion in funded debt, it is implausible in the extreme that they would only have $19 million in payables and other accrued liabilities. In fact, in its last 10-K before filing, RHD showed accounts payable and accrued liabilities of $216 million, 1100% greater than the Disclosure Statement figure.

Taking a case from the pool of low-recovery cases, I looked at Pliant and found it was similar. Pliant listed $93 million of accounts payable on its last 10-K balance sheet  before filing.  But the disclosure statement estimated the general unsecured pool at only $17 million. 

Something has to have happened during those cases to the rest of the accounts payable and similar liabilities.  The discrepancies are so large that further research is warranted into what happened to the rest of the general unsecured liabilities during those cases and the others of the 2009-10 vintages.  I suspect deeper research will show they were mainly taken care of through one or another of the methods I have mentioned above and likely received par or something very close to it.  For example, in Pliant, the debtor was authorized by the court’s order granting the critical vendor motion (docket 47) to pay up to $29 million in critical vendor claims --  almost double the amount of general unsecureds it identified in the disclosure statement.  If  $29 million in unsecured claims received payment in full outside of a plan, and $17 million received 17.5% under the plan, was the case in Pliant, then the unsecured recovery in the case, as opposed to the plan, was 69.5%, a radically different number than Wood’s table shows for that case.

Similarly, Building Materials was a case with recoveries estimated to fall in between Pliant and R.H. Donnelly, so I looked at payouts outside the plan there as well.  Its critical vendor motion was granted for up to $15 million in payments, while its total pool of unsecured creditors, according to the “best interests” analysis done by Peter J. Solomon and appended to the disclosure statement, shows less than $100 million in unsecured claims, so the actual recovery by unsecureds there was substantially higher than the 55% shown in the disclosure statement.

If other cases show similar facts, that would have significant repercussions for the thesis that unsecured creditors are somehow getting a bad deal in chapter 11’s.  It may be that, like drunks looking for their keys under a lamppost (because that’s where the light is), researchers have been looking for general unsecured recoveries in disclosure statements, because that’s where the recovery estimates are published, while in each case the real explanation for the apparent decline in unsecured recoveries may lie elsewhere on the docket.  It may well be the case that much of the purported difference in recoveries between the two samples in Wood’s article can be traced to the greater use of critical vendor and similar motions in the more recent sample.  And persons seeking to improve the lot of unsecured creditors in chapter 11 cases can either rest easier knowing their concerns are misplaced, or simply codify the judicially well-accepted practice of critical vendor and similar first day orders and ensure the result occurs in all districts and not just the most experienced ones.

There are many more avenues for unsecured creditors to be paid in chapter 11:  insurance, letters of credit, cure payments, statutory priority, etc.  While some of the methods for paying unsecured creditors outside a plan reflect choices made by prior bankruptcy legislation, all ought to be taken into account to assess the relative treatment of secured vs. unsecured creditors in the real world as opposed to uninformed academic studies or agenda-driven reform efforts based on statistical garbage as opposed to real fact.

[1]           Bizarrely, Lear is one of the cases for which the UCLA-LoPucki Bankruptcy Research Database contains a figure for unsecured creditors’ recovery (56.4%) that is reasonably accurate  if one blends all unsecured claims and recoveries into one pot.  That the database was right to begin with makes it even harder to understand Wood’s deviation from it.
[2]           In addition to those recoveries, a further $100 million was paid out to unsecured creditors under “first-day” orders as well. It should also be noted, in evaluating the “42%” recovery estimate for the one impaired class, that, according to Lear’s February 14, 2013 press release, “since November 2009 when Lear resumed trading on the New York Stock Exchange following its emergence from bankruptcy … the Company's equity market valuation has more than doubled.”