Bankruptcy verite.

• Author: LoPucki, Lynn M.
• Position: Response to article by James J. White in this issue, p. 691

TABLE OF CONTENTS INTRODUCTION I. THE COMPETING VALUATION METHODS A. Common Aspects of Methodology B. Our Valuation Method C. White's Valuation Method D. Three Errors in White's Valuation Method 1. The Overvaluation of Reorganizing Companies at Filing 2. The Deduction of Current Liabilities from Reorganization Recoveries 3. The Deduction of Cash from Reorganization Value II. SELECTION BIAS III. EXPLANATIONS A. Thin Soup B. Investment Bankers' Reputations C. Auction Process CONCLUSION APPENDIX I didn't do it, no one saw me do it, there's no way you can prove anything! --Bart Simpson (1)

1. INTRODUCTION

   In the empirical study we report in Bankruptcy Fire Sales, (2) we compared the recoveries from the going-concern bankruptcy sales of twenty-five large, public companies with the recoveries from the bankruptcy reorganizations of thirty large, public companies. We found that, controlling for the asset size of the company and its presale or pre-reorganization earnings ("EBITDA"), reorganization recoveries were more than double sale recoveries. (3)

   We are honored that Professor James J. White has chosen to comment on our study. White is an eloquent defender of the status quo, pulls no punches, and always has something interesting to say. Bankruptcy Noir (4) is no exception.

   In his response, White reconstructs portions of our study using some of our data and some data he gathered. He accepts our calculations of the sale prices but otherwise values the companies using a method--Total Enterprise Value--commonly used by investment bankers in valuing companies for takeover. By his use of this method, White has pushed us to situate our findings within the investment bankers' frame of reference.

   White's findings conflict with ours. Examining the same set of companies, White finds no statistically significant difference between sale prices and reorganization values. (5) The conflict, however, results entirely from four errors in White's method. First, White values grossly insolvent companies at filing based on their debts rather than their assets, thus creating billions of dollars in phantom assets. (By phantom assets we mean assets the companies themselves did not claim existed.) Second, in comparing the sale and reorganization recoveries, White deducts current liabilities from the reorganized company recoveries without making the corresponding deduction from the sold company recoveries. Third, after deducting those current liabilities-which are a proxy for cash and other current assets--White deducts the cash a second time. Here too, he makes the deduction from the reorganization recoveries, but not from the sale recoveries. White's fourth error was to drop from his study the seven telecom sale cases with the lowest recoveries. He attempts to justify their removal on the ground that they were unreorganizable. But he retains three higher-recovery telecom sale cases in the study; and he provides no evidence that the companies dropped could not have been reorganized.

   Part I first compares the valuation methods White used with the ones we used and then explains the first three of White's errors. Part II analyzes White's removal of "the telecoms" from his sample and responds to his argument that we should have removed them from ours.

   In his response, White argued that even if our findings were correct, they do not justify our accusations of systemic corruption. In Part III of this reply, we explain why they do. Ultimately, we conclude that White's study is fatally flawed, that his exclusion of the lowest recovery telecom sale cases is unjustified, and that our corruption-based explanation for the disastrous sale recoveries remains the best available.

1. THE COMPETING VALUATION METHODS

   To understand White's valuation errors, one must first understand the valuation methods used by White and ourselves.

   1. Common Aspects of Methodology

      Similarities dominate. Both studies examine the same sample of sixty cases. (6) Both employ the same basic approach to comparing the sale and reorganization processes. That approach is to compute a recovery from each bankruptcy. In sale cases, the recovery is the sale price; in reorganizations, the recovery is the surviving firm's value based on market capitalization. Both studies also determine the companies' values at filing, and calculate the ratio of the recoveries to the values at filing. (7) Both evaluate the sale process by comparing the average ratio of sale recovery to value at filing with the average ratio of reorganization recovery to value at filing. Table 1 shows the common structure of the two studies:

      TABLE 1 COMMON ASPECTS OF VALUATION METHODS Sale Recovery Reorganization Recovery Sale Price Firm value at confirmation Firm value at filing Firm value at filing To investigate further, both studies log all four values to reduce the influence of outliers and use regression analysis to control for differences in the companies' earnings. (8) Both evaluate the sale process by noting whether the sale recovery ratios are statistically significantly lower than the reorganization recovery ratios. (9)

   2. Our Valuation Method

      For the firm's value at filing, we used the "Total Assets" reported by the debtor on Exhibit A of the bankruptcy petition. (10)

      We calculated the Market Capitalization by adding the face amount of the emerging company's debt to the market value of the emerging company's equity. (11) We determined the Sale Price by totaling the consideration paid by the purchaser, whether in the form of cash, the assumption of debt, or the delivery of securities. (12)

      Because our purpose was to isolate that portion of the recovery attributable to the assets existing at filing, we adjusted both the Sale Prices and the Market Capitalizations for substantial changes during the bankruptcy case. For example, if the company borrowed money from a DIP lender after filing and continued to owe it at the time of sale or reorganization, we deducted it from the sale or reorganization value. If the company distributed cash to creditors prior to the sale or reorganization, we counted that cash as part of the recovery. (13)

   3. White's Valuation Method

      White used Total Enterprise Values ("TEV") as the values of all companies at filing and for the reorganized companies at confirmation. (14) For the Sale Price, he used the Sale Price variable we calculated. (15) A comparison of Tables 2 and 3 shows White's substitutions:

      TABLE 3 WHITE VALUATION METHOD Sale Recovery Reorganization Recovery Total enterprise value at Sale Price confirmation Total enterprise value at filing Total enterprise value at filing White defines TEV as equaling "the face value ... of interest-bearing debt, and the market value of common stock." (16) White provides no operational definition of "interest-bearing debt" in Bankruptcy Noir and does not explain the source of his figures. But a comparison of some of his data to the companies' financial statements indicates that he used the amount of long-term debt shown on the companies' balance sheets. (17)

   4. Three Errors in White's Valuation Method

      TEV based on the book value of long-term debt is, in general, an accepted method for valuing solvent companies, including companies emerging from bankruptcy. White has, however, used it improperly in at least two respects. First, he used it to value insolvent companies. Second, he compared TEVs that value only the core assets of reorganized companies to Sale Prices that value both the core and non-core assets of sold companies.

      1. The Overvaluation of Reorganizing Companies at Filing

         White uses the amount of long-term debt shown on the balance sheet plus the market value of equity to value the companies at filing. (18) White provides no examples of scholars or investment bankers who have used balance sheet debt to value insolvent companies and no authorities approving this aspect of his method. His use is erroneous. In discussing enterprise valuation, Stephen Moyer states:

         Debt typically should be valued as shown on the GAAP balance sheet. There are at least three exceptions to this rule.... Finally ... if the firm is in financial distress, then it may be more appropriate to value the debt using market prices as opposed to its balance sheet value. Thus, if there is $200 in debt but it is trading at 40, then it may be more appropriate to value it at $80 instead of $200. (19) To see why balance sheet debt should not be used, assume that a company owes long-term debt of $1 billion, but has no assets whatsoever. White's method would value the nonexistent assets at $1 billion.

         White did essentially that in his valuation of Arch Wireless. On the last financial statement Arch Wireless issued before filing, the company listed assets of $696 million and long-term liabilities of $1,726 million. (20) (White takes his long-term liabilities figure, $1.802 billion from a different source which he does not identify, but the difference is immaterial for current purposes.) (21) Despite the firm's balance sheet negative equity of $1.106 billion (using White's long-term debt figure), White included the entire amount of Arch Wireless' long-term liabilities in his valuation. (22) The effect was to create more than $1 billion in phantom assets.

         Compounding that error, White added another $71 million--the trading value of Arch Wireless's stock. (23) How could the stock of such a grossly insolvent company trade for $71 million at the time of bankruptcy? The answer is that it didn't. White used the last trading value available prior to filing--that value was as of a date seven months prior to filing (24)--apparently before the market knew the depth of Arch Wireless's financial distress. Thus, White valued Arch Wireless--a company that claimed assets of only $696 million (25)--at $1.802 billion plus $71 million, for a total of $1.873 billion. (26)

         Arch Wireless was not an isolated case. The proportion of TEV that was phantom assets--defined as White's TEV value minus...