Bankruptcy contracting revised: a reply to Alan Schwartz's new model.

• Author: LoPucki, Lynn M.
• Position: Response to article by Alan Schwartz in this issue, p. 343

In Bankruptcy Contracting Reviewed,(1) Alan Schwartz purports to restate and defend the bankruptcy contracting model he presented in A Contract Theory Approach to Business Bankruptcy.(2) What he in fact does is abandon key assumptions of the original model and substitute new ones. The resulting new model is driven by reputational constraints neither present nor possible in the original model. Yet it works no better than the original.

The linchpin of Schwartz's response is his insistence that his original model contained an unstated assumption prohibiting debtor firms from lying. In the context of Schwartz's model, the effect of the new assumption is to bar the debtor from strategic behavior at the contracting stage--an implausibility in the context of bankruptcy. To fit this new assumption to his original model, Schwartz had to make new supporting assumptions that he ultimately could not reconcile with the original model.

In addition, as Schwartz has filled in more of the details of his model, other problems of inconsistency and incompleteness have come into sharper focus. As will be apparent, Schwartz has not yet demonstrated the feasibility of bankruptcy contracting in the real world or in his revised model.

1. CREDITORS LENDING AT DIFFERENT TIMES

   Schwartz's original model assumed that only the debtor firm had the information necessary to determine the optimal bankruptcy contract.(3) Schwartz devoted stage one of that model to showing that when all creditors contracted with the debtor firm at the same time, the debtor firm would propose that optimal contract. Schwartz set forth clearly the reason it would do so: "The firm will offer creditors the contract that maximizes the creditors' expected insolvency return because this will maximize the amount the firm can borrow."(4) That is, the debtor firm offered the optimal contract in response to direct economic incentives, not in order to comply with an implicit assumption prohibiting lying.

   At the second stage of his original model, Schwartz relaxed his assumptions to acknowledge that creditors contract with their debtor firms at different times and that the terms of the optimal bankruptcy contract would change over time.(5) His solution at the second stage was a "conversion term" in the debtor firm's contract with each creditor that would change the bankruptcy terms of the creditor's contract to the bankruptcy terms of any contract the debtor firm entered into with a later creditor.(6)

   In my reply to Schwartz, I pointed out that his solution at the second stage destroyed the incentives that had caused the debtor firm to propose the optimal contract in the first stage. An opportunistic debtor firm could understate the amount of the necessary bribe to get a favorable loan from the initial creditor, knowing that it could correct the understatement in its contract with the second creditor. This strategy of understatement would enable the debtor firm to get below-market terms on the initial loan without receiving a lower bribe in the event of bankruptcy. Because the initial creditor would anticipate the strategy and have no way to prevent it, no loan transaction could occur.

   Schwartz does not dispute the devastating effect of the understatement strategy on his model. Instead, he claims that the strategy "is ruled out by [an] assumption"(7) that parties cannot "commit fraud" or "lie to each other at the contracting stage."(8) Schwartz does not claim to have expressed this assumption in his original essay. Instead, he asserts that it was implied because "[c]ontract-theory models assume that parties will not engage in fraud."(9)

   Schwartz cites no authority for that proposition. Nor is it generally true. To the contrary, the introduction to a recently published contract theory text cautions that "[a] homo economicus who possesses private information should be expected to try to manipulate it, since he has in effect a monopoly over his own piece of private information.... The theory of contracts originates in

   these failures of general equilibrium theory."(10)

   Schwartz cites two articles in which the authors expressly assumed that the contracting parties would be truthful,(11) but it is equally easy to cite articles in which the authors expressly assumed the contracting parties in their models could lie.(12) More to the point, the authors Schwartz cites do not make the no-lying assumption merely because it is a convention of contract theory; they make it because it is plausible given the other assumptions of their model.(13) By contrast, Schwartz's original model assumes that the private information (the private benefits) is "unverifiable," making an assumption of truthful disclosure implausible.(14)

   Contrary to the claim he now makes, Schwartz's original essay assumed that debtor firms could lie. To find that assumption in the original essay, it is first necessary to understand the role that the no-lying assumption plays in Schwartz's new model. The lie Schwartz would prohibit is the debtor firm's "contractual promise to choose the optimal bankruptcy system."(15) Debtor firms did not, however, promise to choose the optimal bankruptcy system in Schwartz's original model; they merely "offer[ed] creditors a ... contract that will induce the firm to choose the optimal bankruptcy system in the event of insolvency."(16) Schwartz implies both the promise(17) and its fraudulent nature from the mere fact that the debtor firm contracts for a suboptimal bribe while it has the information necessary to fix an optimal bribe.(18) Because Schwartz implicitly assumes that every debtor firm has that information,(19) it follows that any debtor firm that proposes a suboptimal bankruptcy contract is "lying." The prohibition on lying is a prohibition on proposing suboptimal terms.

   Schwartz's original model contemplated that debtors could propose suboptimal terms. After explaining the operation of his conversion term in his original essay, Schwartz notes that "[a] possible objection to these results is that, as regards the renegotiation-proof contract, the firm would strategically not lower the optimal bribe s* in a contract with the later creditor, though circumstances made a lower bribe efficient."(20) Although that passage alone references a strategy of not contracting with the second creditor at all, Schwartz later expands it to the creditor who does contract with the second creditor:

   The firm would also probably not behave strategically because the behavior can be unprofitable. The gain to the firm from strategic behavior is a higher bankruptcy payoff. The loss stems from the firm's having to offer an inefficient contract to the second creditor, thus not maximizing this creditor's insolvency payoff.(21) Though the debtor firm in this passage knew it was offering a suboptimal bribe(22)--and would have no opportunity to correct it later--Schwartz did not protest the strategy as fraudulent. Consistent with the principles of his original model,(23) he claimed that the conduct would be controlled by his model's pattern of incentives, finally concluding that "the expected costs to the firm of offering later creditors inefficient contracts apparently would often outweigh the gains."(24)

   Schwartz's claim that he intended the no-lying assumption is also undermined by its omission from his extensive, detailed listing of assumptions in the original essay(25) and his simultaneous inclusion of an express assumption against collusion.(26) Collusion is a form of fraud.(27) The exclusion of collusion from the model implies that other forms of fraud are permitted.

   The fact that Schwartz did not make the no-lying assumption in his original model should not prevent him from making it in his new model. But in doing so, he removes the elegant mechanism by which his original model operated--incentives based on the monetary returns in his formulae--and substitutes a new, inelegant mechanism--voluntary full disclosure by the debtor in order to maintain the creditors' good will.

   The original model, by contrast, contained no assumptions making it necessary for the debtor to maintain the good will of its creditors. To introduce them, Schwartz first correctly notes that the understatement strategy requires that the debtor firm first propose a suboptimal contract to the initial creditor and then propose an optimal contract to the last creditor.(28) Schwartz then argues that when the debtor firm proposed the optimal contract (at [t.sup.1]): (1) it would have to justify that change to the initial creditor;(29) (2) the attempted justification would cause a loss of good will with the initial creditor;(30) (3) that loss would disadvantage the debtor firm in a later workout or bankruptcy;(31) and (4) that disadvantage would be sufficient to deter the debtor firm from using the strategy.

   These assumptions are, in effect, the assumptions of a revised model. They are both inconsistent with the original model and implausible. As to the first assumption of the argument, the debtor firm would not have to justify the change in contracts to the initial creditor because the initial creditor would not know the change had occurred until after bankruptcy. Recall that Schwartz's original model set forth no mechanism by which a change in the bribe negotiated with the second...