The New Corporate Web: Tailored Entity Partitions and Creditors’ Selective Enforcement
abstract. Firms have developed sophisticated legal mechanisms that partition assets across some dimensions but not others. The result is a complex web of interconnected affiliates. For example, an asset placed in one legal entity may serve as collateral guaranteeing the debts of another legal entity within the larger corporate group. Conventional accounts of corporate groups cannot explain these tailored partitions. Nor can they explain the increasingly common scenario in which one creditor is the primary lender to all or most of the legal entities in the group.
This Article develops a new theory of selective enforcement to fill these gaps. When a debtor defaults on a loan, the default may signal a failure across the entire firm, or it may signal an asset- or project-specific failure. Tailored partitions give a primary monitoring creditor the option to select either project-specific enforcement or firm-wide enforcement, depending on the signal that the creditor receives. In this way, the creditor can address firm-wide risks and failures globally while locally containing the costly effects of project-specific risks and failures. This option for selective enforcement reduces the costs of monitoring and enforcing loan agreements and, in turn, reduces the debtor’s cost of capital.
These concepts of selective enforcement and tailored partitions have important implications for legal theory and practice. In addition to providing a cohesive justification for the web of entity partitions and cross liabilities that characterize much of corporate structure today, they also inform how bankruptcy courts should approach a wide range of legal and policy issues, including holding-company equity guarantees, good-faith-filing rules, fraudulent transfers, and ipso facto clauses.
author. Assistant Professor of Law, The University of Chicago Law School. I thank Kelli Alces, Ken Ayotte, Adam Badawi, Douglas Baird, Omri Ben-Shahar, Erin Casey, Michael J. Casey, Wendy Epstein, Henry Hansmann, Rich Hynes, Ed Iacobucci, Saul Levmore, Alex Morgan, Ed Morrison, Anthony Niblett, Randy Picker, Eric Posner, Roberta Romano, Andres Sawicki, Alan Schwartz, Julia Simon-Kerr, David Skeel, Holger Spamann, George Triantis, David Weisbach, and Yesha Yadav for helpful comments and discussion. I also thank participants at the Annual Meeting of the American Association of Law and Economics, the Annual Meeting of the Canadian Law and Economics Association, the Canadian Economics Association Annual Conference, the National Business Law Scholars Conference, the Junior Business Law Conference, the Young Bankruptcy Scholars’ Workshop, and workshops at the University of Texas School of Law, Yale Law School, Northwestern University School of Law, and Columbia Law School. I thank Daniel Epstein, Kaitlinn Sliter, and Michael Zarcaro for excellent research assistance. The Jerome F. Kutak Faculty Fund provided research support.
Legal scholars have only a basic understanding of dynamic corporate groups. Existing theories demonstrate that assets of a common economic enterprise might be separated to partition risk,1 create withdrawal rights, or securitize assets in bankruptcy-remote entities.2 But these theories leave unsolved a puzzle created by the various combinations of partitions and overlapping obligations that exist in many large corporate groups today. This Article introduces a theory of tailored partitions and selective enforcement to shed some light on that puzzle. By revealing nuanced motives that drive the specific design of many entity partitions and the contractual relationships that connect them, these concepts move us closer to a cohesive theory that will permit us to understand the modern corporate web that binds assets within an economic enterprise.
Firms regularly separate assets and place them in different legal entities to create value.3 That value may come from risk partitions, withdrawal rights, regulatory compliance, tax planning, or some other source. Existing scholarship often examines these partitions as if firms either fully isolate assets by a legal partition or fully integrate them in one entity. In other words, partitions are considered to be the result of a binary, all-or-nothing decision.4 I argue that this dichotomy is unrealistic and has muddied the theoretical waters.
In reality, firms can tailor the impact and degree of any legal partition to create a precise structure. Certain contract provisions—such as cross guarantees, cross defaults, and holding-company guarantees (collectively, cross-liability provisions)5–can be coupled with legal partitions to create a web of commonly owned assets, targeted liabilities, and precise enforcement options. The prevalence of these tailored partitions is apparent in the capital structures at the core of many recent corporate bankruptcies. The coupling of legal partitions and cross-liability provisions was visible in the bankruptcies of Kodak, Dana Corporation, Calpine, Residential Capital, Visteon, MSR Resorts, and many other corporations.6 In each of these cases, the debtor filed as a group of commonly owned legal entities. For example, while Kodak is a single economic firm, the “Kodak bankruptcy” was actually the administrative consolidation of sixteen different bankruptcy proceedings. Each of the sixteen debtors was its own legal entity, but they were commonly owned, and each entity had cross guaranteed the secured debt of the other entities.7
While bankruptcy proceedings make these structures particularly salient and transparent, we can also observe this corporate structure when large public corporations take on major debt. The public filings associated with those transactions reveal the prevalence of tailored partitions. The secured debt that JCPenney recently took on ($1.85 billion), for example, was cross guaranteed by all of JCPenney Company, Inc.’s domestic subsidiaries.8
I show that these tailored partitions create value by allowing the debtor and its creditors to achieve a balance between specific and general creditor enforcement in response to varying signals of project failures.9 Where two projects are partially but not fully related—say a luxury hotel and a budget hotel—the firm can tailor partitions to allow common risks and failures to be dealt with collectively and to permit independent risks and failures to be addressed in a targeted and contained fashion. The availability of these enforcement options lowers the firm’s cost of capital because creditors can more effectively monitor risk and respond to defaults.10
Recognizing this structural option changes the analysis of corporate groups. Under conventional models, creditors with no specialized expertise loan to the firm as a whole while creditors with expertise focus on particular projects. These models assume that different creditors will specialize in monitoring different projects within one firm.11 But that is not how things look on the ground. It is increasingly common for a single sophisticated creditor12 to monitor both the firm as a whole and the various projects individually. My theory of tailored partitions and selective enforcement can explain this. The central creditor loans to each legal entity while creating cross-liability provisions. When one entity defaults on its loan, the creditor then possesses a valuable selective-enforcement option: it can 1) call a firm-wide default; or 2) selectively waive or ignore some defaults while taking action on others. The second option allows the creditor to focus remedial action on a specific project.
In the budget- and luxury-hotels example, consider a simplified scenario in which the budget hotel’s default sends the sophisticated creditor one of two signals: 1) managers are generally incompetent, and the problems will spread to the luxury hotel; or 2) managers are incompetent only at managing the budget hotel, and the problems will not spread. Tailored partitions give the creditor the option to take action against the entire firm in response to signal one (by way of the cross-liability provisions) or only as to the specific project in response to signal two (by waiving its formal rights under the cross-liability provisions).
The first option—firm-wide enforcement—is valuable because it allows the creditor to act on general signals to contain firm-wide losses. The creditor need not wait for the second hotel to default to assert its enforcement rights. Many of the large bankruptcies mentioned above fit this model. Kodak’s bankruptcy was precipitated by a general demise of its business. While its traditional operations were shrinking because of technological changes in the market, the firm had also failed to move aggressively into new digital markets.13 At the same time, the company was burdened by massive post-employment obligations resulting from a decade of workforce reduction.14 Its potentially profitable licensing business was stalled in litigation with the likes of Apple, RIM, and HTC.15 Throw in an unprecedented financial crisis, and it is not surprising that the bankruptcy of Kodak included all of its domestic entities.16 One can safely assume that the primary creditors would have prevented any restructuring efforts that did not address all operations.
The second option—project-specific enforcement—is valuable because it reduces the significant ancillary effects caused by firm-wide responses to project-specific problems17 and prevents other parties from opportunistically forcing the default to spread. Examples of these project-specific enforcement actions are less common in the bankruptcy dockets because one of the benefits of contained enforcement is that it allows the primary creditor to avoid bankruptcy proceedings altogether. When a creditor can limit the hold-up power held by others, it can push for an out-of-court restructuring more effectively. Examples of creditors opting for project-specific enforcement are therefore more likely to take the form of waived guarantees in enforcement actions. For example, Sunstone Hotels let ten of its forty-two hotels go into default18 when it experienced financial trouble. Because of a cross-default provision in its bond indenture, this move gave the bondholders the right to call a firm-wide default that would have likely collapsed the entire enterprise into bankruptcy. The bondholders opted to forgo the firm-wide enforcement option and voted to amend the indenture to remove the threat of cross default. This allowed Sunstone (and its bondholders) to walk away from ten hotels (including the W in San Diego) without triggering the rights of any other creditors on the thirty-two remaining properties, including Hilton, Marriott, and Renaissance hotels across the country.19
As I demonstrate below, firm-wide enforcement is not available when there are partitions without cross liability, and project-specific enforcement is not available without legal partitions. Intuitively, one might think that project-specific enforcement could be achieved through security interests. But in a world with multiple creditors, this is not the case.20
The failure to recognize that tailored partitions create these valuable options causes confusion in the courts and introduces unnecessary puzzles and complexities. For example, some scholars hold the view that corporations undo the entire effect of entity partitioning by causing affiliated legal entities to agree to cross-liability provisions.21 These scholars wonder why a corporation would partition an entity just to re-integrate it at the next moment. Why create a corporate web when the firm could just partition or not partition?22
The concepts of tailored partitions and selective enforcement dislodge this riddle and reveal major implications for bankruptcy law. By examining how these structures create value, I attempt to provide some guidance on difficult questions surrounding issues like fraudulent transfers, ipso facto clauses, and good-faith filing in bankruptcy. In Part I, I describe cross-liability provisions and tailored partitions and explore why it has been difficult to fit them into existing theories of asset partitions. In Part II, I demonstrate how tailored partitions and selective enforcement work and how they create value. I also describe some of the more common variations on the structures used to create enforcement options. In Part III, I examine the implications that these theories of tailored partitions and selective enforcement have for the laws of finance and capital structure, focusing primarily on bankruptcy law. In Part IV, I discuss potential critiques of these theories.
I. cross liabilities, the corporate web, and conventional theories of asset partitions
The coupling of entity partitions with contractual cross liabilities provides a broad variation of capital structures from which debtors23 can choose.24 I do not attempt to catalog all of those possibilities here. I do, however, provide a prototypical example. In this Part, I describe some common components of the cross liabilities and discuss how the practice of combining cross liabilities with asset partitions compares to conventional accounts of corporate structure. In the Appendix, I have included the language of some common contract provisions that the parties use.
A. Cross Liabilities
Most commonly, the cross liabilities at play will be cross defaults or cross guarantees. The cross-default provisions with which I am concerned run across entities. Intra-entity guarantees are less puzzling. For example, a large primary loan25 to one entity might have a provision that treats the default on any other debt to that same entity as a default on the primary loan. The reasoning behind this structure is straightforward. The default on one obligation is a signal of distress that a primary creditor wants to take into account in monitoring that debtor: it is the canary in the coal mine.
The puzzle posed by the corporate web and the analysis of tailored partitions arises only when the default crosses legal boundaries. The inter-entity cross default will cause a loan to one entity to be in default whenever an affiliated entity defaults on a debt obligation. To illustrate, Bank may make a large loan to SubCo. It will then include a provision that states that if AffiliateCo defaults on any “material indebtedness,”26 SubCo is in default on its loan from Bank. The default by AffiliateCo does not have to be on a loan from Bank. It can be on any material loan. Thus, if AffiliateCo misses a payment on any major debt obligation—a loan or a major supply contract, for example—that might trigger a default on SubCo’s loan from Bank.
An example of this structure can be found in the $750 million unsecured credit facility that Darden Restaurants, Inc. (owner of the Olive Garden restaurants and the former owner of Red Lobster) took out in 2007.27 That agreement provided that Darden would be in default on the credit facility if “any Material Subsidiary (i) fails to make any payment . . . in respect of any Material Indebtedness” or defaults in any other way.28
A cross guarantee29 is similar to a cross-default provision. But the default of one entity does not necessarily trigger the default on any other loans. Instead, the cross guarantee makes the guarantor entity liable for the default of the direct borrower. If AffiliateCo borrows $1 billion, and SubCo guarantees it, then AffiliateCo’s default puts SubCo on the hook for $1 billion.
Cross guarantees will often run in both directions. So SubCo and AffiliateCo might collectively borrow $2 billion, and each might guarantee the other’s obligations. The cross guarantees may also be coupled with cross-default clauses. In that way, the default of either entity on any loan (not just the $2 billion primary loans) would be a default of both entities on the $2 billion.30
Notably, the cross guarantees are almost always “unconditional guarantees of payment” and not “guarantees of collection.” This means that the creditor can go after the guarantor for payment without taking any action against the primary debtor. In the large corporate context, where these loans might be in the hundreds of millions or billions of dollars, this point is particularly important because a default will often trigger the lender’s right to accelerate the loan. That means the remaining balance of the loan becomes due immediately. If the primary debtor defaults because it missed a payment, the lender then has the option to go straight to the guarantor entity for payment in full. This option will generally give the lender the power to foreclose on the guarantor’s assets or force it into bankruptcy without taking any action against the primary debtor.31
In the context of large corporate credit facilities, the exact structure of the guarantee will vary. The credit facility might call for joint-and-several liability of the various legal entities. Such facility would provide an open line of credit that designated entities could draw upon. Later, it would provide that the designated entities “jointly and severally, hereby absolutely, unconditionally and irrevocably guarantee the punctual payment [of the debt] when due . . . , and all obligations of each other Loan Party and each other Subsidiary of the Company now or hereafter existing under or in respect of the Loan Documents.”32 The parties designated as jointly and severally liable might or might not be identical to the parties designated as borrowers who can draw on the line of credit. Alternatively, the guarantee might be set forth in a separate guarantee agreement executed by all guarantor entities.33
Cross liabilities are common when large corporations (public or private) take on debt through a primary creditor.34 With the other possibility—multiple creditors holding multiple options—the bargaining dynamics become more complicated, since one lender can destroy the option of another. The analysis in this Article suggests, then, that selective enforcement creates the most value when one major creditor, syndicate,35 or other unified group possesses the option. Any of these forms can function as a primary creditor.36 Thus, we should expect to see the selective-enforcement option appear most frequently in the hands of a single creditor or group of creditors who negotiate ex ante agreements with each other to coordinate their behavior. In fact, we do see coordination between creditors over the use of various enforcement options in publicly available loan documents.37 To enhance the value of their selective-enforcement option, primary or major banks lending to corporate groups also include provisions limiting the use of selective-enforcement options by other creditors. For example, a review of loan documents for large publicly traded corporations reveals that many contain provisions prohibiting the debtor from putting cross-guarantee provisions in agreements with other lenders.38 When we see multiple lender groups that do have cross-guarantee provisions in their loan documents, there is often a corresponding inter-creditor agreement that coordinates the use of those provisions.39 For example, it is common for a first lien lender to demand that junior debt (second lien or unsecured credit facilities or junior notes) be subject to a standstill agreement that prohibits junior creditors from taking actions to enforce defaults without permission from a senior lender for a set period of time.40
These various cross-liability arrangements interact with legal partitions to provide creditors with an ex post choice to invoke asset partitions in response to some risks or to ignore them in response to others. Below I explore the way in which this selective-enforcement option creates value. But first, we have to understand why assets are partitioned at all. In the next section, I explore the conventional account of asset partitions.
B. The Conventional Models of Asset Partitions
The separation of unrelated risks is the most commonly identified goal of asset partitions.41 By this account, a firm will partition unrelated assets to separate the risks associated with them. Related assets, on the other hand, will be kept together in one legal entity. Choosing the right structure reduces the cost of capital by improving creditors’ ability to monitor the debtor.
As a preliminary matter, it is worth noting—as it will come up later—that the existing literature often conflates the ideas of enforcement and monitoring. In the literature, monitoring assumes the ability to enforce.42 For purposes of my analysis, however, it is necessary to separate the two. Monitoring will refer to oversight intended to detect signals of value loss. Enforcement will refer to action taken in response to those signals. I show below that the need for specific enforcement rights is the driving force behind tailored partitions.
In the risk-partition model, integrating related assets creates value.43 Two oil refineries in Texas, for example, can be monitored by one creditor with expertise in the region and the industry.44 Some have suggested that this creates economies of scale for monitoring.45 The point is far from obvious. A creditor can ignore separate legal entities if it wants to create the economies associated with integration. All it needs to do is contractually require that the debtor provide consolidated financials.46
But integration provides other cost savings. For example, full integration eliminates administrative and management costs associated with maintaining separate legal entities. Likewise, integration creates less paperwork for certain transactions: one loan document is less expensive to write than two. Moreover, there are economies of enforcement. It is cheaper for a creditor to conduct one rather than two enforcement actions (such as a foreclosure or bankruptcy proceeding). Finally, and most importantly, the law often restricts enforcement options to only one project if the projects are not integrated. A default at Refinery A alone does not trigger enforcement rights against Refinery B if they are not integrated—even when the default of A reveals information that B is failing as well. All of the savings here arise when the assets and their default risks are related in some way.
Things are different when the assets are unrelated. Consider a firm that owns both an oil refinery in Texas and a hotel in New York. Those projects are more costly to finance if the firm places them in one legal entity.47 Imagine that the firm has two primary unsecured creditors and owes each the same amount. One creditor specializes in monitoring oil refineries and the other specializes in monitoring hotels. If the hotel assets become worthless without either creditor’s detection, the two creditors will be left fighting over the oil refinery’s assets as protection for their investments. In this example, the creditor that monitors oil refineries will have lost value because of the other creditor’s failure to monitor the hotel.48
Though the literature generally speaks about this scenario in terms of specialized “monitoring,” the real driver is enforcement. A creditor can always require a debtor to keep separate books and records for different assets even without a legal entity partition. This option allows the creditor to monitor assets separately just as if there were an entity partition. But the creditor has little incentive to do that when all enforcement measures bleed across assets.
This bleeding will always occur when the assets are housed in the same legal entity. A creditor who specializes in monitoring the oil refinery has to enforce against the firm as a whole when it receives a signal. For example, any bankruptcy proceeding will include all of the firm’s assets. Even a foreclosure sale of one asset will implicate the rights of the creditors of the firm as a whole if they claim that the sale is below the asset’s value or that it has certain other adverse consequences. As Ed Iacobucci and George Triantis point out, all enforcement actions will be taken against a legal entity.49 So without an entity partition, there is no way to fully contain an enforcement action against a single asset or group of assets. The outcome of that enforcement action will therefore depend on the combined condition of both the oil refinery and the hotel.
As a result, the integrated firm has a blended capital structure that compromises asset-specific financing. Because the failure of any one asset will ripple across the entire firm, a creditor cannot contain enforcement to the failing asset—even with asset-specific security interests.50 Creditors must, therefore, monitor (and charge for monitoring) risk in both the energy and travel industries or charge a premium for assuming a risk they cannot monitor and respond to effectively. The crucial point is that the assets have different risks that are not correlated, and the monitoring expertise to reduce those risks lies with different lenders.51
This blending will also increase the cost of credit if—as most people assume—different capital structures produce different monitoring incentives that are optimal for different assets.52 For example, riskier projects are less likely to be financed with public debt. Likewise, unproven management may need to adopt a structure that includes expert monitors with security interests.53 A blended capital structure prevents this risk tailoring. Additionally, managers of integrated firms can more easily cross-subsidize between projects to serve private interests, which will lead to a higher cost of credit. This opportunity to divert value exists unless the debtor can credibly commit not to take advantage of it by using asset partitions.54
C. Limitations on the Conventional Model
The discussion up to this point presents a choice—integrate or partition. And the optimal corporate structure depends upon whether the assets are related or unrelated. The implicit assumption is that partitions and asset relationships are binary. This assumption makes the point salient and the models elegant. A single creditor can monitor related assets as one bundle. Separate creditors can monitor unrelated assets in separate legal entities. But when the relationship between assets is not all or nothing, the optimal partition is not all or nothing either.
Moreover, firms often employ partitions even when there is a single creditor monitoring all projects. These are not cases where creditors specialize in different projects, but rather where one creditor is monitoring all assets that are nonetheless divided into different legal entities. Indeed, the common characteristic that has appeared repeatedly in the bankruptcies, out-of-court restructurings, and public loan documents of the last decade is a primary creditor sitting above the entire firm. This is true for Kodak,55 JCPenney,56 Sunstone Hotels,57 Smuckers,58 and others. For these cases, the analysis of tailored partitions and selective enforcement provides insight into the corporate web.
To understand selective enforcement, we need to think not about hotels and refineries but about assets that are differentiated in more nuanced ways. For instance, a luxury hotel and an economy hotel may experience the same value loss if the real estate market crashes. But they may be affected differently by a general economic downturn. A strain on income of wealthy travelers, to take another example, may benefit a budget hotel at the expense of the luxury hotel. The optimal enforcement response to a signal indicating a real estate market crash differs from the optimal response to a signal indicating a general economic downturn. As a result, a capital structure that allows the creditor to choose the response ex post is more valuable than one that locks in the response ex ante. In the next Part, I examine a hypothetical capital structure to show how the mechanisms of tailored partitions and selective enforcement provide creditors with the option to select the optimal ex post response and how that option creates value.
II. tailored partitions and selective enforcement
To summarize what is to come: the demonstrative example for tailored partitions will include two assets whose performances are closely but not completely correlated. A primary creditor59 financing these assets will face a dynamic enforcement project. This creditor can monitor some aspects of management and risk jointly in a bundle. Some signals produced from those monitoring efforts will pertain to the entire firm. They will carry information about the future performance of both assets. But other aspects of risk and management of the two assets will be unrelated. Because information about those risks will be limited to a single asset, enforcement mechanisms will, in turn, be optimally contained to that single asset.
The two assets in my example are a luxury hotel on Chicago’s lakeshore and a budget hotel near Chicago’s O’Hare Airport. Many aspects of managing these assets are related, but some are not. In each period, a primary creditor monitoring these assets receives one of three signals for each asset. For the luxury hotel: 1) no signal; 2) management is incompetent at everything; or 3) management is incompetent just at running the luxury hotel. Because signal two suggests firm-wide incompetence that will spread to the management of other assets, the primary creditor will want to react by calling a default that can be enforced against both hotels. For signal three, on the other hand, the primary creditor will want to contain the default to allow it to take enforcement action60 against the luxury hotel while allowing business at the budget hotel to continue as normal.
Put differently, the failure signals provide information about the expected return and value of each loan. Signal two tells the primary creditor that both loans have lower expected returns. As long as the creditor has other investment opportunities, it will want to take cash out and invest in better projects. Signal three tells the creditor that only the luxury-hotel loan has fallen in value. In response to this signal, the creditor will want to cash out on only the luxury-hotel loan. There is no reason to cash out on the budget-hotel loan if it still has returns at or above market.
The following sections examine a firm and its capital structure in various scenarios. I look at 1) a firm with perfectly correlated risks and operational characteristics across projects, to demonstrate the value of a legal integration; 2) a firm with perfectly uncorrelated risks and characteristics across projects, to demonstrate the value of a legal partition; and 3) a firm with partial correlation and partially related characteristics across projects, to demonstrate the value of tailored partitions and selective enforcement.
A. Selective Enforcement: A Simplified Example
Consider an entrepreneur (“Entrepreneur”) who has identified a property on Chicago’s lakefront for developing a high-end hotel. The hotel will have great views and access to the major attractions of the city. Entrepreneur forms HotelCo, a Delaware corporation of which she is the sole owner. HotelCo has received approval from the city to build the hotel. The only thing Entrepreneur needs is financing. But this is not a problem. She has a strong record of accomplishment in the luxury-hotel business, so banks have lined up to lend to HotelCo.
We will start with a single-creditor structure (which, for now, makes it unnecessary to discuss security interests and priority). HotelCo borrows $1 billion from a bank (“Bank”)—which knows the hotel industry well—to finance the project, and things go well. As is often the case, the success of one project leads to another, and Entrepreneur decides to expand. She has three options: 1) build another luxury hotel on Chicago’s lakefront; 2) build an oil refinery in Texas; or 3) build an economy hotel near Chicago’s O’Hare airport.
When she approaches Bank, the lending officer (being familiar with the legal scholarship on risk partitioning) knows exactly what to say to options one and two but is mystified by how to deal with option three.61
1. Option 1: Perfect (or High) Correlation and Integration
Because option one is the conventional case of perfectly (or nearly perfectly) correlated assets, Bank is happy to finance the new luxury hotel and suggests that the project be called LuxuryTwo and be wrapped into the HotelCo legal entity. Bank will simply double the loan, and both hotels will serve as collateral62 for the entire loan. The fact that these are separate hotels contained in one corporate entity is of no import because Bank is expert at monitoring loans to luxury hotels, and it has precise covenants in place to measure Entrepreneur’s performance in running these hotels. At the first sign of incompetence, Bank will call a default that allows it to foreclose on both hotels, push the entire entity into bankruptcy, or force Entrepreneur to take certain actions with respect to the entire enterprise. Similarly, Bank has its thumb on the pulse of Chicago’s hotel industry and real-estate market. Once again, at the first sign of a decline in either market, Bank will use its covenants to impose control over the entire operation and likely force the sale of both hotels.63 Recall that we start with the extreme assumption that long-run performances of the hotels are perfectly correlated. There will be no failure at one property without failure at another.
Under these conditions, neither Bank nor Entrepreneur wants to create a separate legal entity for LuxuryTwo because the partition introduces unnecessary costs. For Entrepreneur, the partition creates administrative costs that have no value to the business as a whole. These costs are likely small in a relative sense, but they are also unnecessary. With the two hotels integrated into HotelCo, Entrepreneur can have one loan agreement, one management team, one tax return, one bank account, one payroll, and so on. Similarly, HotelCo can enter into common contracts more easily. And in the event of a major corporate event, Entrepreneur can sell the company as a unit and avoid the renegotiation of various contracts.64 The integrated firm with its financing structure is depicted in Figure 1.
Beyond the elimination of administrative and transaction costs, the benefits of legal integration are largely under-theorized. Scholars have generally accepted integration as the status quo and focused on justifying deviations from it. In passing, some suggest that information economies are the primary value of integration.65 But the existence of those economies is questionable.
Bank has likely achieved economies of scale in information by lending to hotels throughout the city regardless of legal partitions. For example, Bank could choose to investigate only one hotel to which it lends and use that information as its signal for further action across all loans. Even if information came out at different times for different borrowers, this strategy might be a cost-saving mechanism for Bank. Why monitor ten properties when you need to monitor only one?66 Regardless of the source of the information, the corporate structure of LuxuryOne and LuxuryTwo will have no effect on whether Bank gets the information. Bank need only investigate the hotel and real estate market once under any structure. Bank may get even better general information about the hotel industry (and LuxuryTwo) from monitoring LuxuryOne. Similarly, looking at just LuxuryOne and LuxuryTwo, with identical risks and operational characteristics, Bank need only assess Entrepreneur’s competence once.67 But, again, that information is available regardless of the partition.
Counterintuitively, some have suggested that the information economies from integration arise not when assets are similar but when they are uncorrelated and the investor is seeking diversification.68 This suggestion is unpersuasive. Perhaps the blending of information reduces monitoring, when one assumes that a debtor’s manager is better at compiling a diversified portfolio and monitoring it than the creditors. Most empirical work, however, suggests the opposite.69 And in any event, information blending can easily be required in a partitioned firm. The legal partition changes neither the blending nor the expertise in diversification. Both can be implemented regardless of the corporate structure as long as the same manager is at the top of the hierarchy. The only cost benefits of integration here are the reduced administrative costs in maintaining fewer legal entities and documenting fewer loans.
Another argument for integration when assets are unrelated may be that integration provides deeper protection for a creditor. For example, consider a creditor (“Creditor”) that loans $100 to project A and $100 to project B. There are no other assets. If A loses 10 and B gains 10, enforcement against an integrated firm leaves the creditor whole. Enforcement against separate legal entities, on the other hand, leaves creditor down ten dollars. Basic finance theory tells us that this alone is not a convincing justification for integration. After all, the gain in reduced risk to creditor is achieved only by shifting that risk to the equity holders. The cost of capital remains constant.70 Put another way, the lender could achieve the exact same expected outcome by loaning to separate entities and then hedging with an investment in equity.
Buried in and implicitly conflated with the idea of information economies is a major source of real value that integration provides for creditors: economies of enforcement. Monitoring and information economies do not require integration, but enforcement economies do. Returning to perfectly correlated assets, Bank knows that any sign of distress at either hotel indicates distress at the other as well. But the defaults that allow Bank to act on these signals may not occur simultaneously. Perhaps Bank has taken advantage of monitoring efficiencies involved in lending to related projects (for example, by doing inspections on only one property). Bank may have observed and documented a default on one project. But documenting and verifying the same default on the other project will be costly and time-consuming. Alternatively, default triggers may simply materialize at different times.71
Bank, therefore, does better when the assets are integrated in one entity. Bank will include a long list of default triggers in a loan agreement. Sometimes these defaults serve as early signals that something may be wrong.72 But sometimes the defaults are just noise—technical violations that provide no signal. For example, a borrower may be late in providing some financial information required under the covenants. The delay may be the result of the borrower’s scramble to solve a major problem or it may be just an administrative oversight.73 The initial default leads to an investigation. If it turns out that the business is doing fine and the default was unimportant, Bank will often waive the default. If the investigation shows problems, Bank will take action. Bank may waive default in exchange for renegotiated terms or (in the extreme case) call a default and accelerate the debt. Other times, the defaults are simply technical violations that allow Bank to act on non-default signals. Bank may have received a non-default signal and have been waiting for a default to act.74
If Bank knows that the real estate market is taking a dive, it will want to enforce against both LuxuryOne and LuxuryTwo. When the properties are integrated into HotelCo, Bank needs to wait only until a default occurs with regard to one of the projects. That default allows Bank to exercise its rights firm-wide as to both projects. If, on the other hand, LuxuryOne and LuxuryTwo are partitioned into separate legal entities without cross liabilities, Bank would have to wait for two defaults to exercise full control. LuxuryOne, Inc.75 might default first, making it clear to Bank that it needs to call default on both loans. But in the absence of a default by LuxuryTwo, Inc., the legal partition would keep Bank from doing this. In the face of perfect or high correlation, waiting for LuxuryTwo, Inc. to default before taking action will force Bank to sit on the sidelines knowing that assets are wasting away. The delay also gives management time to take costly gambles with LuxuryTwo. Those gambles can have negative expected total returns but positive expected private returns to Entrepreneur.76 Bank will, therefore, not want LuxuryOne and LuxuryTwo to be partitioned. And Entrepreneur will want to use the integration ex ante to lower her cost of credit. The lower cost of credit results because the integration creates a de facto commitment to not take gambles with Bank’s money. In the absence of Entrepreneur’s ability to signal that she is a type who never gambles opportunistically,77 this commitment is valuable.78
Crucial to the analysis below, these economies of enforcement can be achieved through an alternative mechanism. Where assets are partitioned, Bank can manufacture the same enforcement rights by demanding contractual cross-liability provisions. Therefore, even if LuxuryOne and LuxuryTwo are owned by separate entities, the default by one can trigger enforcement rights against both.
Still, for perfectly correlated assets, there is no enforcement difference between integration and partition with cross liability. There is hence no reason to incur even the minimal additional administrative costs of partitioning assets and the transaction costs of negotiating cross-liability provisions. Additionally, under current law the enforceability of cross-liability provisions is far less than certain.79 Integration is therefore the optimal mechanism to create enforcement economies when there is no independent reason to use a partition.
Sometimes, however, reasons unrelated to risk may require partitions. For example, a firm doing business in multiple jurisdictions might create separate legal entities to ease compliance with different regulations and tax regimes.80 Because the regulatory benefits of partitioning can be significant and the administrative costs of partitions and cross-liability provisions discussed above are likely to be relatively small, we should expect firms to combine partitions and cross liabilities to create tailored partitions that achieve compliance savings while maintaining the value of enforcement economies.
These cases present the strongest case for cross liabilities and the corporate web. Bank will insist on cross-liability provisions that undo the economic impact of the artificial regulation-driven legal partition. The result is a legal partition for regulatory or other purposes with an economic re-integration by way of cross-liability provisions. As long as the re-integration is not frustrating some other valuable end—such as regulatory compliance goals81—the cross-liability provision should be viewed favorably.82 The focus of the remainder of this Part is on more nuanced motivations.
The takeaway so far is that firms will rarely partition perfectly correlated assets for risk allocation, monitoring, or enforcement purposes. But as the administrative and transaction costs are likely to be small, and few assets are actually perfectly correlated, it is not surprising that firms (for the reasons discussed in the next two sections) routinely partition and agree to cross-liability provisions.
2. Option 2: No Correlation and the Benefits of Partitions
Consider another example, in which Bank is asked to finance a hotel and an oil refinery. Again, the analysis of this case tracks conventional accounts. There is little to no correlation between the hotel and the oil refinery. Bank will not only refuse to finance the second project, but it will also demand that Entrepreneur partition the second project into a second legal entity. Bank only has expertise in monitoring hotels. If Entrepreneur builds a refinery, finances it by another lender, and houses it within HotelCo, then the success of Bank’s original investment will turn in large part on the success of the refinery. Bank has limited ability to monitor this project; as a result, it would have charged a much higher interest rate for the loan in the first place if HotelCo reserved the right to enter the refinery business. This is an extreme example of ex post risk alteration.83 Almost certainly, Bank will have required—in the original loan agreement—a covenant prohibiting HotelCo from undertaking the refinery project without Bank’s blessing. And HotelCo would have likely agreed to that covenant to get the best interest rate.84 So Entrepreneur creates HotelCo, Inc. and OilCo, Inc.
By partitioning the assets, Entrepreneur can go to a different lender who specializes in monitoring oil refineries to get a loan. This structure is depicted in Figure 2:
The new lender wants nothing to do with the hotel business. And with a partition, neither lender worries about the other project. A failure of OilCo will have no effect on HotelCo. Bank can focus on monitoring the hotel industry, and the lender to OilCo, Inc. can focus on monitoring the refining industry. Because these operations are completely unrelated, none of the enforcement economies discussed above will be available, and nothing is lost from the partition. It would be surprising if we saw integration or cross-liability provisions adopted in this case.
Indeed, the use of cross-liability provisions in cases of no correlation should be the most suspect of all. Such provisions suggest inefficient cross subsidies that are 1) a sign of incompetence; 2) opportunistic risk alteration; or 3) the use of internal capital markets to transfer value and circumvent the limits that external capital markets have placed on one of the projects.85
Each item on this list should raise concerns, but the third problem is the most concerning. With respect to the first problem, incompetent subsidies destroy value, but any transfer or transaction that a business undertakes in its day-to-day operations may be the result of incompetence. Rarely does the law provide a mechanism to second-guess business transactions based on incompetence alone.86 The second problem, risk alteration, is potentially costly because creditors usually loan to an entity based on its assets and the expected risk and return from future uses of those assets. When Entity A takes on liabilities for the debts of Entity B, the expected risk and return change. A is taking on B’s liabilities and reducing B’s exposure to risk. Even if A receives a market premium compensating it for the new exposure, it has just changed the risk profile of its business. Because a market premium was paid, this does not change the relative value of the firm as a whole, but it does change the value of the relative investments across the different layers of investors. At the same time, risk alteration is unavoidable and occurs every day with every loan. Any purchase, sale, payment, or other transfer can change the firm’s risk profile to some degree. The possibility of some risk alteration is, therefore, priced into every loan or, where possible, the risk alteration will be prohibited by covenant.
Additionally, if A does not receive a premium payment for those liabilities, then A has just transferred to B the market value of the risk reduction. The lenders to A are now protected by a smaller set of assets (and the lenders to B are protected by a larger set).87 This is the equivalent of a dividend or cash give away. The law plainly allows such transfers in the absence of fraud or constructive fraud. Again, the risk of such transactions is priced into every loan.
The third problem—subsidies to circumvent external market limitations—is different. Such subsidies are rarely possible unless management is hiding movements in internal capital markets to achieve artificially low costs in the external market. That is, they are possible where creditors are being tricked and think they are lending on a safe project when they are actually lending on a much riskier project. This destroys value for lenders and creates a misallocation of capital in the economy as a whole.
Because the first two possibilities—risk alteration or incompetence—are at best value-neutral motivations, and the third is always value-destroying, courts should be especially skeptical in cases in which cross liabilities are incurred between entities with uncorrelated risks. We may then desire fraudulent transfer law to be relatively strong here. Fraudulent transfer laws invalidate certain transactions that either harm creditors or make it more difficult for them to monitor assets.88 Because there is no value-creating justification here, prohibiting these transactions might be favored. Rules that prohibit value-destroying transactions without affecting value creation are the gold standard. This point suggests a strong justification for the importance that the law of constructive fraudulent transfer places on the “for value” requirement in these cases.89 If the transaction includes a true market premium for the cross liability, then there is no value transfer from one entity to the other. Entity B has not circumvented the costs of external markets because it paid the market price to Entity A.90
A market premium, however, is not likely to have been paid. In our example, HotelCo, Inc. is rarely going to be the lowest cost lender to OilCo, Inc. A guarantee from HotelCo, Inc. to OilCo, Inc. would be unnecessarily costly. Imagine that the finance market would charge OilCo, Inc. $100 for a loan guarantee from an outside investor (“Investor”). Investor, the guarantor, is liquid, diversified, and has expertise in monitoring refineries. The same guarantee from HotelCo, Inc. should cost more (say $110). This is because HotelCo, Inc. has a higher cost of capital than Investor has and is not as effective at monitoring. The differential will materialize as an increase in the capital cost for HotelCo, Inc. Now that HotelCo, Inc. is on the hook for OilCo, Inc.’s debts and it and its creditors cannot efficiently monitor OilCo’s operations, investors in HotelCo, Inc. will charge a higher interest rate to the tune of $110.91 In our example, then, either HotelCo or OilCo loses out. OilCo, Inc. pays $110 for a guarantee that it could have received for $100. Or OilCo, Inc. pays $100 to HotelCo, Inc. for a guarantee that costs HotelCo, Inc. $110. In both cases, HotelCo, Inc. incurs an increased cost of capital of $110. Entrepreneur’s total enterprise is out $10 relative to a market transaction. Thus, even if OilCo, Inc. pays a market premium, the transfer is still likely to be a sign of incompetence or of opportunistic risk alteration.
In this situation, courts often place a heavy burden on OilCo, Inc. (or its creditors) to show that it paid equivalent value for the guarantee when HotelCo, Inc. is insolvent and the guarantee results in a loss to HotelCo, Inc.’s creditors.92 This may be an effective rule. But we might consider an even heavier burden, such as treating the lack of correlation as an independent badge of fraud, one that could substitute for insolvency. Of course, the transaction would not be automatically invalid. Generally, the law requires two badges of fraud to invalidate a transfer under the law of constructive fraudulent transfer.93 Thus, a cross guarantee with no correlation would be valid if it were for value and invalid if not.
On the other hand, there is one scenario in which cross liabilities of this sort may be value creating and appropriate: where there is private information that cannot be conveyed to the market. The joint managers of the conglomerate may have information about the success of the projects available to OilCo, Inc. but may not be able to convince any lender of the accuracy of that information.94 In that case, if HotelCo, Inc. is sitting on uninvested cash or available credit, the best investment may be the one about which it has inside information. Because these cases should be rare,95 the aggregate cost of invalidating this small subset of valuable transactions may be relatively small.
The bigger caveat is that a court may find it challenging to create a clear metric for measuring correlation. This is no small matter. Correlation is a fuzzy thing, and the best entrepreneurs are likely to see correlation where others do not. Indeed, the logic above might be taken to mean that courts should second guess all decisions companies make to enter new markets or that the law should require those companies to use legal partitions without cross liabilities when they enter those markets. After all, expansion into a new market has a similar impact on creditors. But the idea of a court deciding that a computer company like Apple should not have entered the telecom industry is worrying. Short of a meaningful and objective metric for correlation, we should be reluctant to change significantly the law in this context.
In the end, this is a dynamic problem. Correlation is often indeterminable. Any rule that turns on correlation will, therefore, be difficult to operationalize. But this analysis suggests that the absence of correlation is an evidentiary factor that should be weighed in close cases when other indications of inefficient risk shifting are present.
3. Option 3: Partial Correlation, Tailored Partitions, and Selective Enforcement
In this section, I explore considerations that arise when the firm is looking to go forward with two partially related projects. Entrepreneur’s two projects will be LuxuryOne (her luxury hotel) and EconoRoom (her budget hotel). The risks facing a luxury hotel on the lakeshore and a budget hotel near the airport are correlated across some dimensions but quite distinct across others. Conventional accounts imply that there is a binary switch at some point along the continuum. For a lot of correlation, integrate; for very little, partition. This assumes that in a world of partial correlation, the parties must simply accept the second best. The parties bear the costs of partitioning if they are less than the costs of integration, and vice versa.
But this ignores the tools available to lenders and borrowers to create value in structuring deals. If it were possible to tailor the partition, lenders could reserve the option to respond to firm-wide signals96 globally and uncorrelated project-specific signals locally on a case-by-case basis.
There are several dimensions across which risk can be partially correlated. In the hotels example, these might include real estate markets, luxury- and economy-hotel markets, geographic hotel markets, and Entrepreneur’s skill at managing the two types of businesses. While the real estate near O’Hare and on the lakeshore will be equally affected by the general economy in Chicago, a dramatic shift in crime near downtown Chicago might affect only the real estate value of LuxuryOne. Similarly, the impact of shifts in the tourism business will be highly correlated between the projects. But a decline in local tourism (vacationers from the suburbs) may affect LuxuryOne without affecting EconoRoom. And while Entrepreneur may be an astute businesswoman with a knack for property management, her experience with luxury hotels may not translate into success with the budget traveler staying next to the airport.
I focus on Entrepreneur’s management skills. Assume that Bank still has the expertise that it takes to monitor both projects. This is not like the oil refinery business. But Bank’s monitoring will produce different signals about the business in different states of the world. Assume that Bank looks at the cash flows and operation reports of a hotel to know whether management is doing its job. The borrower has agreed to provide accurate books and records to Bank on a quarterly basis. When Bank receives reports from LuxuryOne, it receives one of three signals: 1) no new information; 2) management is incompetent at everything; or 3) management is incompetent at just the luxury-hotel business. When it receives reports from EconoRoom, it receives a similar set of signals: 1) no new information; 2) management is incompetent at everything; or 3) management is incompetent at just the budget-hotel business. As discussed above, signals two and three may trigger defaults in some cases.97 In other cases, the firm will look for a default to act on an unrelated signal.
As a starting point, Bank’s desire to limit the flow of internal capital markets is not on its own a compelling explanation for partitions here.98 An internal-capital-markets theory would suggest that a large bank may finance all projects of the firm but require the debtor to create entity partitions simply to restrict the flow of capital between projects. The limitation on this story is that the cross-liability provisions are a direct path for capital to flow from one project-entity to another. LuxuryOne, Inc. might borrows funds secured by the assets of EconoRoom, Inc. When it does so that is the functional equivalent of moving capital from EconoRoom, Inc. to LuxuryOne, Inc.
EconoRoom, Inc. will have reduced borrowing capacity, and LuxuryOne, Inc. will have increased borrowing capacity. Indeed, parties often draft credit agreements to allow any entity to draw on the total amount of the revolving loan. The loan may be for $1 billion to be allocated among the entities as the debtor sees fit. In those cases, the lenders do not restrict the flow of capital between the project-entities. If anything, the provisions allow the capital to flow at the managers’ discretion while maintaining the lender’s option to choose which asset to enforce against.99
Some have suggested that the partitions serve Bank’s interest in getting accurate and separate books and records for each project. Under certain circumstances, Entrepreneur may have an incentive to obscure signals about her incompetence on a given project. Secretly moving assets from one project to another may accomplish this smokescreen. It will be at least marginally easier to do this under a one-entity structure.100 Commingling funds within one legal entity will likely be easier to defend ex post as mere incompetence rather than outright fraud. In a partitioned firm, assets can technically be smuggled across legal boundaries, but there will be more hoops for Entrepreneur to intentionally jump through in order to create a smokescreen. Those hoops make it easier to verify the fraudulent intent.
Still, I find this point to be weak support for the claim that partially correlated assets are partitioned to deter fraud. The hoops implicit in a partition can be created (at least roughly) by covenants. Thus, entity partitioning created to deter fraud will, in most cases, be unnecessary. And when fraud is occurring, it is not at all clear that the extra hoops created by legal partitioning will do much to deter someone who has already accepted the more significant expected threshold costs of committing fraud.
More important to Bank than the relatively rare case of fraud will be what it can do with the signals of non-fraud risk that it receives. If Bank receives signal one from both projects (no new information), it does nothing. It does not matter how the firm structured its partitions. If Bank receives signal two from both projects (firm-wide incompetence), it enforces against both projects.
But Bank’s response becomes more complicated when it receives signal two from LuxuryOne and signal one from EconoRoom. If the assets are integrated in one legal entity, Bank will enforce against both assets. Signal two tells it as much about EconoRoom as about LuxuryOne, and the value-maximizing response is to call all the loans before the incompetence worsens or spreads. This is not possible when the assets are fully partitioned into separate legal entities. Without cross liability, signal two is a default by LuxuryOne, while signal one is not a default by EconoRoom. In this scenario, Bank must sit on the information that EconoRoom is about to crash until it defaults separately. Bank cannot preemptively intervene the way it could if the entities were integrated. In the meantime, EconoRoom is depreciating in value.
Management and equity will have strong incentives to take on self-interested, risky projects that have negative expected value for LuxuryOne as a whole. To see why, consider a firm that is likely to fail. Managers have to make a choice about how to use the firm’s remaining assets. If they take a conservative approach, the firm can be wound down with some value left for the creditors (but none for equity). Now let’s say that there is also a risky alternative that will destroy the firm’s entire value ninety-nine times out of one hundred but has a one-in-a-hundred chance of creating a huge payout that will save the business. The managers and the equity-holders get the upside of the risky project and bear none of its downside. They prefer a 99% chance of total failure to a 100% chance of moderate failure.
Ex post, Bank in this scenario wishes that the entities were integrated. But the solution is not for Bank to demand integration ex ante. To see why, consider what happens when Bank receives signal one from LuxuryOne (no new information) and signal three from EconoRoom (project-specific incompetence). This situation essentially captures the world in which Entrepreneur has proved to be a successful manager of luxury hotels but an incompetent manager of budget hotels. Here, Bank wants to enforce only against EconoRoom and leave Entrepreneur to run LuxuryOne.
The following example demonstrates this intuition. Let’s say that when Bank made both loans, it expected a 10% return (adjusted for risk) from each. Assume also that Bank has limited capital and that if it had more capital it could take advantage of other investment opportunities with a 9% expected return. A year has passed, and Bank has received a signal of failure at EconoRoom but not LuxuryOne. Now the expected return on the remaining value of EconoRoom is 5%; LuxuryOne still has an expected return of 10%. The rational action for Bank is to call a default, and cash out only the loan to EconoRoom, and reinvest the money at 9%. Any recovery against LuxuryOne would lead to a loss of 1% in expected returns. Trying to recover losses on the EconoRoom loan from the value of LuxuryOne would be akin to the sunk-cost fallacy. The best that Bank could do is reinvest the money recovered from LuxuryOne in the market at 9%. When Bank calls a default against LuxuryOne or an integrated entity that owns it, however, the debtor might even be able to use a bankruptcy proceeding to force a renegotiation of the loan at somewhere between 9% and 10%—even when the loan prohibits prepayment or refinancing!101
The facts as I have assumed them to be until now—with no other creditors—might still allow Bank to limit its actions to just the EconoRoom assets in a world of no partitioning. Bank could call a default, take a lien, and threaten foreclosure on some or all assets. Firms with one creditor cannot generally file for bankruptcy, and so Entrepreneur would not be able to do much more than threaten state-court litigation.
But few firms have one creditor,102 and multiple creditors complicate things. If Bank is unsecured, any enforcement action will be against the legal entity as a whole.103 This will trigger the rights of Entrepreneur and the other creditors to initiate firm-wide litigation or to push the entire firm into bankruptcy.
And if Bank itself wants to use bankruptcy remedies, it will have to do so against the entire legal entity. This raises numerous problems for Bank. Bank wants LuxuryOne to continue to operate but wants to liquidate EconoRoom. As noted above, Entrepreneur may use the bankruptcy to refinance the loan on LuxuryOne. Additionally, bankruptcy triggers Entrepreneur’s exclusive control of the plan of reorganization.104 She may be able to use that power to extract value from Bank. Even worse, the general creditors of HotelCo can become obstacles to the restructuring by raising objections to any plan supported by Bank.105
These rights might threaten the value of the primary creditor’s investment in the non-distressed projects and allow third parties (or the debtor itself) to extort some of the investment’s value. In its efforts to keep LuxuryOne afloat, Bank may be forced to make concessions to the other parties who do business only with EconoRoom. On the other side, the general creditors or other parties who do business exclusively with LuxuryOne may now have bankruptcy objections or strategic hold-up options that would not otherwise exist.
A more streamlined process could be accomplished if the bankruptcy involved only EconoRoom and let LuxuryOne continue to operate as usual. In that scenario, the general creditors of EconoRoom could try to extract value from Bank, but their actions would be limited to EconoRoom. The legal partition would facilitate this.
Additionally, Bank may wish to enforce its rights without dealing with a bankruptcy proceeding. But in fully integrated firms, when one creditor calls a default, the enforcement action will often trigger defaults on agreements the debtor has with other parties. These are not “cross-defaults” per se. Rather, the impact of the enforcement action makes it impossible for the debtor to fulfill other unrelatedobligations. For example, a primary creditor may accelerate a large loan forcing immediate payment. Or a secured creditor may sweep the debtor’s cash collateral. Either measure ties up the debtor’s cash flow. This lack of liquidity will cause defaults throughout the entire legal entity. In this way, Bank’s action in response to the failure of EconoRoom provides general creditors with the possibility of pushing all of HotelCo into default and bankruptcy. Entrepreneur may even respond to Bank’s notice of default in ways that encourage other creditors to push for bankruptcy.106 In this way, both Entrepreneur and general creditors can make Bank’s enforcement more complicated. The consequence is leverage for value extraction. This can impose costs on Bank, which must either limit its enforcement options or allow the value to be extracted.
In contrast, the value-extracting strategic maneuvers available to Entrepreneur and general creditors are fewer when the entities are legally partitioned. When Bank threatens to enforce against EconoRoom by liquidating it or taking it over, the best Entrepreneur or general creditors can do is threaten a bankruptcy filing vis-à-vis EconoRoom, Inc.107 This threat imposes costs associated with bankruptcy, but the value of LuxuryOne is unaffected. Moreover, the negotiation and litigation costs of that smaller bankruptcy may be significantly lower. This is particularly true when the additional creditors of LuxuryOne would otherwise complicate the bargaining dynamic.
In some cases, Bank itself may prefer to push EconoRoom into bankruptcy. The primary motive here will be to achieve a free-and-clear sale.108 When the assets are partitioned, Bank can respond to the default signal from EconoRoom by finding a buyer, pushing EconoRoom into bankruptcy, and orchestrating a free-and-clear sale. The court order that accompanies such a sale provides significant value over a foreclosure sale outside of bankruptcy.109
A free-and-clear sale is not so simple when there is one integrated legal entity because the sale can be accomplished only in bankruptcy. But legal entities, not assets, file for bankruptcy.110 Thus, the only way to sell EconoRoom free and clear is to take HotelCo into bankruptcy, with LuxuryOne along for the ride. The various procedural hold-up maneuvers and costs of bankruptcy once again emerge. Partitioning allows Bank to avoid that problem.111 Finally, bankruptcy is much simpler when the assets are confined to separate legal entities.
To summarize, when Bank receives a firm-wide signal, it will want to enforce against the entire enterprise. Cross-liability provisions allow this even if the firm’s assets are partitioned across various legal entities. When Bank receives a project-specific signal, however, it will want to enforce against the failing project alone because broader enforcement will impose costs by triggering hold-up rights for other creditors. Partitions (even with cross-liability provisions) protect the option to enforce on a project-specific basis. Figure 3 depicts the tailored partition:
In short, combining asset partitions and cross-liability provisions allows for precise ex post balancing. The key here is that the boundaries of the partition can be defined after the relevant circumstances have materialized. In the all-or-nothing view, it is a choice of the lesser of two evils. With tailored partitions, the partition can be calibrated to achieve the best of both worlds.
Notably, the theory described above works only when few (one, maybe two) creditors or creditor groups have the option for tailored enforcement. If many creditors have cross-liability provisions, any project-specific default would trigger cross-enforcement rights in the hands of many parties. This reintroduces hold-up opportunities, and the selective-enforcement option would be worthless. Consistent with this, a primary creditor often prohibits the debtor from including cross-liability provisions in loans or agreements with other creditors, or it restricts those provisions to immaterial loans.112
4. An Aside About Security Interests
I have mostly bracketed security interests up until now. One might expect a discussion of security interests here as a possible mechanism for creating traditional partitions or selective-enforcement options. That is only partly true. Security interests do separate priority rights in assets. And they do create asset-specific foreclosure rights. Combining that partitioning with cross liabilities allows for some, but not all, of the selective-enforcement benefits of tailored partitioning.
Cross-liability provisions in security agreements can trigger firm-wide enforcement rights. In the integrated world, default on payment of a loan secured by the LuxuryOne assets can certainly cause a default on a separate loan secured by the EconoRoom assets if the loan agreement provides for that. In that sense, the lender has the choice whether to call default against LuxuryOne or against both hotels. But on the other side of the equation, there are significant limitations to Bank’s ability to contain its enforcement action to just one asset in the absence of a legal partition.
As Iacobucci and Triantis point out, the true value of a legal partition can be traced largely to the “legal personality” of the corporate entity.113 The key is that enforcement actions are taken against legal persons. Entity partitions limit the impact of enforcement actions by defining the boundaries of the legal person against whom actions are taken. Security interests, which are asset specific, do not do this.114
Thus, a security interest in the LuxuryOne assets creates specific priority rights in those assets. Bank may view its security interest in LuxuryOne as providing it with the flexibility to take action against that project alone. But the creditor must take any legal action (other than foreclosure) against HotelCo as a whole. The most extreme example of such legal action is bankruptcy. Bank may want to push LuxuryOne into bankruptcy to achieve a free-and-clear sale. But assets cannot be put into bankruptcy, only legal persons. In a structure with security interests but no legal partitions, any bankruptcy involving LuxuryOne will by necessity involve EconoRoom.115 This triggers the potential hold-up rights for other creditors and Entrepreneur.
Perhaps Bank can get around this by foreclosing on the asset. Still, this limits its enforcement rights to that one remedy. And in any event, even a foreclosure of the LuxuryOne assets may trigger some rights of other creditors of HotelCo. The effects of these other rights—essentially freezing liquidity—can ripple out and cause default on HotelCo, Inc.’s agreements involving EconoRoom assets. The general creditors116 might then push HotelCo, Inc. into a costly bankruptcy. Entrepreneur and other general creditors117 can then use the procedural levers of bankruptcy to extract value from Bank.118 Filing rules are broad enough that it will be quite difficult for Bank to challenge a bankruptcy filing of HotelCo, Inc. in the face of its foreclosure threat and the presence of other creditors. Bankruptcy may reduce the value of LuxuryOne for all involved, or it may simply impose transaction costs on Bank. In turn, the threat of filing may alone be enough to extract value from Bank when Bank threatens foreclosure on EconoRoom. Additionally, if Bank forecloses and sells LuxuryOne, any other general creditor might later argue that the sale was a fraudulent transfer.
Security interests and entity partition will, therefore, generally have different goals: the former is targeted at priority of payment, and the latter at enforcement options.119 Because they achieve different goals, security interests and partitions are often paired together. A lender who desires priority may also desire an enforcement option.120
B. Variations on a Common Theme: Holding Company Guarantees and Subordinated Primary Creditors
The analysis above presents simplified examples of affiliate cross guarantees. But the actual capital structures we see in practice can have layers upon layers of entities. Firms can combine overlapping webs of cross-liabilities of varying form to choose very specific enforcement rights. I pause here to identify two noteworthy variations on the model presented above.
1. Holding-Company Guarantees and Stock Pledges
One common structure is a guarantee of subsidiary debt from a parent holding company whose only asset is its equity in the subsidiaries. This is depicted in:
equity guarantees/stock pledge
Again, under existing accounts one might puzzle at this structure. Why has the debtor created all of these separate entities only to cross guarantee all of the partitions away? Moreover, what value is a security interest in equity of the subsidiaries? After all, such a security interest is usually the second- or third-most junior possible position in the capital structure of an enterprise (the most junior being equity in the holding company). These questions can all be answered by analyzing the selective-enforcement rights that are at work.
Imagine one likely signal is that management is incompetent but that the assets are fine.121 When management is incompetent, Bank wants to oust them from every entity. In our previous examples, that would require calling defaults or cross defaults at every entity. But calling cross defaults between the operating companies carries the risk of triggering hold-up rights for all other creditors.
This is where the equity guarantee comes in. By combining the equity guarantee with the operating companies’ cross guarantees, the structure allows for all of the options discussed above: calling a default on Subsidiary A; calling a default on Subsidiary B; or calling a default on both subsidiaries. But the equity guarantee provides another option: calling a default on the parent holding company (“Parent”) alone. The effect in this case is an immediate foreclosure right on the equity of the subsidiaries. The lender then has a right to appoint new directors of all of the operating entities and take complete control of operations. Indeed, under some agreements the pledged stock documents will be delivered to the lender at the time of the loan, along with executed instruments of transfer, irrevocable proxies, and an acknowledgement of equity interest registration page.122 The lenders draft these documents with the purpose of transferring control and voting rights (and allowing the replacement of directors) instantaneously upon Parent’s default or notice of default.
When the equity guarantee is invoked by calling a default only against the parent, the primary creditor does not get any liquidation right over the subsidiaries’ assets. But the guarantee does not trigger any rights of general creditors of the operating subsidiaries.123 The operating companies continue with business as usual, but with the primary creditor calling all of the shots.
If the assets were wasting away, the lender would use its other guarantees to call a default and liquidate the assets. But when the signal is simply that the company needs new management, the lender can take over cleanly and run or sell the company. This provides a much different option for enforcement than a simple security interest or cross guarantee. Consistent with this analysis, lenders often include a covenant that explicitly prohibits the parent company from having any other creditors. Indeed, the covenants will go as far as prohibiting the company from doing anything at all (beyond the administrative tasks required to exist as a holding company).124
The temporary restraining order in Madison Capital Funding LLC v. HomeOrganizers, Inc.125 illustrates how these provisions are intended to work but also how uncertainty about courts’ responses can impair their effectiveness. In that case, Madison Capital, the primary creditor, had made loans to HomeOrganizers, Inc. and its subsidiaries. HomeOrganizers, Inc. was the holding company parent for various operating companies, including Closet World, CBD Franchising, Inc., Home Closets, Inc., Closets By Design, Inc., Closet World Arizona, LLC, Closet Dimensions, Inc., and CBD Las Vegas, LLC (collectively the “operating subsidiaries”).126 The loans were guaranteed by all entities and were secured by all of their assets, including the equity that HomeOrganizers, Inc. held in all of the operating subsidiaries.127 After the HomeOrganizers group had defaulted on several covenants, Madison Capital sent a notice to all entities containing the following: 1) notice of ongoing default; 2) notice of its exercise of voting rights; 3) written consents showing its votes to replace all of the subsidiaries’ directors with a sole director chosen by the creditor; and 4) notice of the express instruction that no officer shall take any action outside of the ordinary course of business without approval from the new director.128
Madison Capital is consistent with the above-described model of selective enforcement against a holding company. The assets of the operating companies did not appear to be losing value, so firm-wide enforcement was not desirable. At the time of the default notice, the operating companies were, in fact, experiencing record sales.129 Instead, the creditor’s actions were aimed at taking control to oust management or sell the business without involving any other creditors.
The old managers of the HomeOrganizers Group130 responded to the default notice by threatening to put the subsidiary entities into bankruptcy. Madison Capital then immediately brought an action in state court seeking a restraining order and injunction to prevent the old management from filing for bankruptcy.131Soon afterward, the court issued a temporary restraining order.132 The judge likely reasoned that under the loan documents, HomeOrganizers was now under the control of Madison Capital.133 And so the old managers had no authority to put HomeOrganizers into bankruptcy.
At this point, from a state-law perspective, Madison Capital owned and controlled the operating subsidiaries.134 No other creditors had any rights triggered by foreclosure, so the only remedy available to the old managers should have been a state-law action arguing that the foreclosure of the equity in the operating subsidiaries was not proper under the loan agreement. If Madison Capital prevailed on that issue, it was in the driver’s seat to continue to run the companies, or hold a foreclosure sale and auction the equity in the operating companies to the highest bidder. In such a sale, Madison Capital would have been entitled to the proceeds up to the full amount it was owed, including any default or prepayment penalties.
The old managers of HomeOrganizers, Inc. took a different view. They went ahead with the bankruptcy filing anyway—in direct violation of the state-court restraining order.135 They also filed an adversary proceeding in the bankruptcy court seeking an injunction prohibiting Madison Capital from exercising any voting rights that interfered with the governance of HomeOrganizers, Inc. or the operating subsidiaries.136 Madison Capital, for its part, filed an emergency motion to dismiss the bankruptcy cases.137 This set the ground for a potential jurisdictional battle between the state court and the bankruptcy court. If Madison Capital’s selective-enforcement rights were to be respected, the bankruptcy filing was improper and derogated their state-law contract rights. But if the bankruptcy court were inclined to hear the case, could it be bound by the state-court order? The order was based on the premise that Madison Capital controlled the board of HomeOrganizers. That is plainly a state-law issue, but the order was not a final judgment on the merits. Rather it was a statement of the likelihood of the outcome for the purposes of an ex parte decision on a temporary restraining order. If the bankruptcy court adjudicated the issue first and reached the opposite conclusion, things would have become messy.
As is often the case with uncertainty of this sort, the parties quickly reached a settlement.138 That agreement allowed the old managers to maintain control subject to close oversight and a tight timetable for a sale of all assets of the HomeOrganizers group to satisfy the debt owed to Madison Capital. In the end, the company was sold in bankruptcy (to a fund that also included management), and Madison Capital was paid the full amount owed on the loans (though it is not clear whether they collected any default penalties).139 The use of the selective-enforcement option here functioned to give Madison Capital leverage to achieve a strong settlement, but the legal uncertainty behind the enforceability of the holding-company guarantee left them with less than complete victory of quick foreclosure and sale without a bankruptcy proceeding.
The single holding company model just described is merely the tip of the iceberg. Tailoring can create multiple levels with various iterations and options from which the primary creditor can choose. For example, the structure in Figure 5 might be used.
The loan proceeds would go to the three operating companies with cross guarantees, cross-default provisions, or stock pledges from all entities. Upon receiving a signal and a default trigger from Operating Co. 1, for example, the primary creditor could choose to 1) enforce against Operating Co. 1’s assets; 2) enforce against all the operating companies’ assets; 3) enforce against the equity in Operating Co. 1 to take control of it; or 4) enforce against the equity in all three Subsidiary Hold Co.’s to take control of the enterprise. Even further, if the default of Operating Co. 1 was a signal about Operating Co. 2 and not Operating Co. 3, the primary creditor could enforce only against the assets of 1 and 2 or the equity of 1 and 2, and so on.
2. Subordinated Primary Creditors
The implicit assumption so far has been that the primary creditor holding the selective-enforcement rights is not subordinated to other lenders. Thus, the primary creditor could be a senior secured creditor or an unsecured creditor of a debtor without secured debt. That need not be the case.
A primary creditor can exercise an effective option for selective-enforcement even when some other lenders have senior claims on assets of the firm. The Sunstone Hotels example mentioned in the Introduction provides an illustration of this variation.140
Prior to becoming publicly traded, Sunstone Hotels had a primary secured creditor holding recourse debt that was cross-collateralized by its various hotel properties. Around 2005, Sunstone Hotels began a capital restructuring that included an initial public offering of equity and the renegotiation of its credit facilities. As part of that restructuring, Sunstone also issued public debt in the form of unsecured notes that included cross-default provisions. At the same time, it began transitioning its secured debt to separate non-recourse mortgage loans on each hotel that was not cross-collateralized.141
The result of the change was that the secured lender could respond only to a default on a project-specific basis. A payment default on one hotel allowed the secured lender to foreclose on that specific hotel but not on any of the others. This was true regardless of the fact that the same bank held the loans to each hotel. At the same time, the cross-default provisions in the unsecured bonds entitled the bondholders to call a firm-wide default any time one hotel defaulted on a payment to any lender (including a default on the secured loans). Thus, when ten hotels went into default in 2008 and 2009, the secured lender had a threat of foreclosure on each hotel.142 But the threat stopped there. Only the bondholders could decide whether to opt for enforcing the defaults against the remaining thirty-two hotels. In this case, the majority of bondholders consented to an amendment to the bond indenture that removed the threat of cross default.143 The bondholders’ consent to amendment suggests they believed that the demise of the ten hotels was not likely to spread.
This is all to say that the primary creditor holding the selective-enforcement option does not need to be the most senior lender in the capital structure. The factors that dictate which lender will play the role of selective enforcer at the center of the web are complicated. The likelihood that a secured lender will be over-secured on a given project, the volatility and option value associated with a given project, and the particular monitoring expertise of the available lenders all influence who will be the selective enforcer. An over-secured lender on a volatile project, for example, may be too quick to pull the default trigger at the expense of more junior creditors. A general lender with no industry expertise may be too slow to pull the trigger and thus not willing to take on the central monitoring role.
The availability of relevant market information likely also plays a role. A senior secured lender might be the best central monitor for a privately held company when the information necessary to assess firm-wide risk has to come from compiling information gathered from active, close monitoring of all assets. For a public company with publicly traded debt, on the other hand, the securities markets provide a constant flow of information about assets. Analysts who specialize in watching the hotel industry will transfer information into price changes for publicly traded shares and bonds in real time. In those cases, the role of the secured lender in assessing firm-wide risk (as opposed to asset-specific risk) may be less important. The key now is that a centralized (potentially unsecured) authority has the power to respond to information signals from the market, not that a lender has the expertise to monitor and produce those signals. One might therefore predict that a company like Sunstone would do exactly what it did here: shift to a subordinated primary creditor structure after its equity became publicly traded.
At its core, the inquiry into why a firm might choose a specific lender as its primary or major creditor is simply an extension of the existing theory that different risk profiles demand different optimal capital structures. Finance literature has long pointed out that different types of creditors make better monitors for different types of assets.144 I add only that different types of creditors also make better primary enforcersfor different types of assets.
III. implications for law and theory
A theory of tailored partitions and selective enforcement provides insight into increasingly complex and varied corporate structures to understand what motivates the agreements that form today’s corporate webs and how the law should treat these webs. In this Part, I explore the implications of the foregoing analysis for some of the pressing legal questions posed by the new corporate web.145
A. Good-Faith Filing
For the selective-enforcement option to have value, it must be real. Much rests, therefore, on the rules of good-faith filing. The takeaway then is that when the primary creditor opts to enforce against one entity, the courts should keep the other entities separate. An enforcement action against Affiliate A should not trigger the right of a completely stable Affiliate B to file for bankruptcy. The law should not allow the junior creditors or equity holders of either entity to use the default event of A alone to justify the bankruptcy filing of B. The Bankruptcy Code, however, has vague rules on when a court can dismiss a case as improperly filed; these rules give vast discretion to the judge.146 One implication of the analysis above is that it is problematic when the courts do not police filings to ensure that the legal entity (rather than the corporate group) has met the good-faith-filing requirement.147 This is why the primary creditor will often prohibit other loans from having cross-default provisions in the first place.148
A preferable standard that preserves the selective-enforcement option would require the court to look at each debtor separately. Just as a debtor should not be allowed to file for bankruptcy to benefit an unrelated third party, it should not be able to file to benefit an affiliate that is a separate legal entity. Moreover, the standard should be more precise than “good faith.” A solvent debtor filing for bankruptcy should be required to articulate a non-affiliate based story on how the bankruptcy process will enhance the value of that entity for the benefit of its creditors. And it should be required to do so in detail with supporting financial documentation.
B. Fraudulent Transfer Law
Every cross-liability provision results in a value transfer and a risk alteration. But transfers and alteration occur with virtually every transaction a firm engages in. The prototypical transfer is a dividend. Dividends do not create value ex post. They just transfer money from the firm to equity holders. That said, dividends may add value ex ante by providing a mechanism for equity to recognize returns on their investment. This can potentially lower the cost of capital. In any event, the law plainly allows dividends, though they can be abused. They can, for example, shift value to equity in a way that creates a value-destroying transfer and risk alteration for creditors. Either that risk is priced into loans, or the creditors demand covenants to prohibit dividends.
Cross-liability provisions create a value transfer that is less likely to be abused than dividends. Indeed, most negative value transfers that might be created by a cross-liability provision can be manufactured more cheaply by way of a dividend. Cross-liability provisions transfer value by creating a liability at one entity for obligations of another. This requires a three-party transaction with fairly specific terms. Dividends on the other hand create minimal transactions costs and involve nothing more than a direct transfer of value. The upshot is that a prohibition on cross-liability provisions is likely to destroy value. As long as dividends are legal—and they almost certainly will continue to be—a prohibition on cross-liability provisions will not reduce the opportunistic transfers that can be achieved by dividends, but it will destroy the value-creating option of selective enforcement.
To put it another way, allowing these partitions is costless in a world in which we already allow dividends. It is unclear why anyone would use cross-liability provisions to transfer value when a dividend is much cheaper. The only opportunity for abuse would occur if cross liabilities were allowed in some cases in which other transfers such as dividends were not. The key is therefore to treat them the same. For the most part, fraudulent transfer law accomplishes this aim.
The law of constructive fraudulent transfer voids transfers made when a firm is insolvent that are not for true value. True value is difficult to measure because there is no market information to compare. This provides firms with the opportunity to hide a dividend or some other transfer in a mispriced premium. If one firm pays an affiliate firm a premium for a cross-guarantee provision, that premium could be too high or too low. If it is too high, that is a transfer of value to the firm providing the guarantee. If it is too low, it is a transfer to the firm receiving the guarantee. If the transferor is insolvent, this is a fraudulent transfer that may be difficult to detect.149
All of this suggests that courts should use the insolvency rule as their main guide. Courts should be skeptical of the value of premiums paid during insolvency. They also should not care at all about premiums paid during solvency. This is essentially the current state of the law. Cases often turn entirely on the insolvency question.
If insolvency is shown, the courts place (or should place) a high burden on transferees to show that the value of the premium was paid.150 To be clear, the black-letter law does not create a special burden of proof on the transferee here. But I am suggesting that courts sometimes impose one—and that they should.151 Doing so would make it costly to adopt selective-enforcement provisions when a firm is insolvent; but that is true of many bankruptcy laws. The risk of misbehavior increases as capital diminishes. The law draws a line somewhere to restrict discretion when misbehavior becomes significantly likely. The line is not perfect, but insolvency is often a good measure, and it is better than the alternatives.152 It is easier to measure than other lines as long as we have a reasonable definition. While the word “insolvency” means different things in different places, here we should use a technical meaning. We need a bright-line rule and do not want ex post judging. The key is for the line to be the same across the board. If the dividend rule were different from the cross-liability rule, then that would create opportunities to structure around rules.153
There is one additional takeaway: since we are treating cross liabilities the same as other fraudulent transfers, we need to examine the law on savings clauses. Constructive fraudulent transfer law states that transfers not for value will be deemed fraudulent if made by an insolvent entity.154 The law goes further to say that if the transfer itself makes the entity insolvent, then it will also be deemed fraudulent.155 Thus, if a cross guarantee makes a solvent entity insolvent, the guarantee can be voided.156
To protect against fraudulent transfer claims where guarantee might render a guarantor insolvent, parties often include a clause—a savings clause—stating that if the transfer renders the debtor insolvent, it will be deemed to be only for the largest amount that would not render the debtor insolvent.157These clauses are common in cross-guarantee provisions, but one recent court decision has cast some doubt on their validity.158 The bankruptcy court in In re TOUSA, Inc., suggested with very little analysis that these clauses were void. That conclusion is difficult to square with a broad view of cross liabilities and fraudulent transfers.159
Take, for instance, the way that dividend payments can be structured to avoid rendering a debtor insolvent. Imagine that a company issued several incremental dividends throughout a day. All the dividends that were paid before the insolvency materialized would be upheld; all those paid after would be voidable. The analysis in In re TOUSA, Inc., implies that guarantees should be treated differently. But there is no clear reason for that distinction. Savings clauses just draw the line at the exact point at which the parties have priced the ex ante risk of transfer. Again, we cannot change that price by prohibiting savings clauses. If the true purpose of the cross guarantee were to transfer wealth, a debtor could simply make a transfer dollar by dollar with dividends until it was insolvent.160 Therefore, savings clauses do not increase the incentives for parties to use cross guarantees as means to opportunistically transfer wealth relative to dividends.
The downside of prohibiting savings clauses is that doing so will introduce great uncertainty and cost to the lending process.161 Insolvency can be difficult to measure precisely, and a court’s measure of insolvency can be difficult to predict with the precision that is required in a world without savings clauses. The same is true of predicting a monetary value for a guarantee that is by its nature contingent.162 Because invalidation is absolute, a mistake of one cent (that is, guaranteeing one cent that transitions the guarantor from solvency to insolvency) would change the entire cost dynamic of the deal. Prohibiting savings clauses on cross-liability provisions would then require parties to do extreme diligence and to hire expensive experts to document their analysis of the potential guarantor’s solvency. This is not worth it because there is no value gained when dividends can be used so cheaply. Put differently, prohibiting savings clauses on cross-liability provisions still allows the company to issue value-destroying dividends cheaply but would effectively eliminate value-creating cross liabilities. This approach is backwards.
C. Bankruptcy and Ipso Facto Clauses
Clauses triggering a cross default from one entity to another raise questions about the Bankruptcy Code’s prohibition on ipso facto clauses. Such clauses, which change a creditor’s rights against a debtor based on the debtor’s filing bankruptcy, are prohibited in some circumstances.163 Examples of such clauses might include a contract term attempting to change the priority for payments to creditors if the borrower files for bankruptcy,164 or a term that shortens or terminates the duration of the contract upon filing.
But what if a bankruptcy filing triggers a change in the rights of an affiliated but separate legal entity? For instance, the bankruptcy of Affiliate A might shorten or terminate the duration of a separate agreement to which Affiliate B is a party; or A’s filing might accelerate the payment or change the priority of payments on a separate loan to B. On their face, these terms would not appear to be prohibited ipso facto clauses. In theory, Affiliate B need not be in bankruptcy. Where that is true, the bankruptcy court has no power over agreements between B and its creditors. The provision should, therefore, be enforceable if state law allows it.
Similarly, if Affiliate B were truly a separate firm owned by an outsider, then the law would certainly respect those clauses. It might therefore be thought that such a clause is valid even when A and B are affiliates in bankruptcy together. One bankruptcy judge, James Peck, has opined on this matter three times in major cases and come to the opposite conclusion.165 Elsewhere, I have suggested that that outcome might not be justified.166 Approaching the issue with the new corporate web in mind, however, reveals support for Judge Peck’s position.
In Charter Communications,for example,the secured lender wanted to invoke the bankruptcy filing of Debtor A in order to irreversibly accelerate hundreds of millions of dollars of debt of an otherwise solvent affiliate, Debtor B. Debtor B wanted to reinstate (and decelerate) its loan. The Bankruptcy Code allows for this if Debtor B can cure all defaults that were not triggered by ipso facto clauses.167 The parties agreed that the default—Debtor A’s filing—could not be cured. The question was whether the default was triggered by an ipso facto clause.168
Abstracting from the doctrinal dispute, it is worth noting that acceleration of affiliate debt is a sure sign that the lender has gone down the path of firm-wide enforcement. Acceleration is a way to get out of an investment and, unless it is for a trivial amount, it will lead to a restructuring or liquidation of some sort. The lender in Charter Communications did not want to see Debtor B survive and continue making payments. It was calling due hundreds of millions of dollars of debt from all entities and then claiming that because of the entity partition, the lender could avoid reinstatement of its investment. If the entities had been integrated, the lender could not have avoided reinstatement based on an ipso facto clause. The lender was essentially treating the entities as integrated for all purposes other than the Code’s prohibition on ipso facto clauses.
This form of strategic positioning in Charter Communications counsels in favor of a finely crafted rule that protects selective enforcement without providing an opportunity for creditors and debtors to use partitions just to avoid the application of the Bankruptcy Code. If we absolutely respect the partition, any lender, knowing that it will always opt to enforce against the entire firm, can use cross-default triggers across entities in place of a prohibited ipso facto clause within one entity. A lender can loan to entities A and B, have each entity cross guarantee the loan of the other entity, and then include a term that changes priority or other rights in A when B files for bankruptcy (and the other way around). When A and B file for bankruptcy together, this is functionally no different from having one large entity with a prohibited ipso facto clause in place. As long as the prohibition on ipso facto clauses is appropriate,169 this is troubling. The rule is avoided, and doing so creates unnecessary transaction costs—much like a tax shelter.
On the other hand, clauses that can be exercised consistent with project-specific enforcement create real value, and the law should respect them. For example, when A is a critical supplier to its affiliate B, it would be common to see a clause that changes the terms of a loan or other agreement to which B is a party upon A’s bankruptcy filing. A lender might increase the interest rate on B to account for the added risk that it may lose its critical supplier. Or a counterparty to another agreement might demand additional assurances of the future performance of B’s obligations. It would be appropriate to respect those clauses when Affiliate B is not forced into bankruptcy with A but not to enforce them when B is forced into bankruptcy. Once B is in bankruptcy, local enforcement is no longer an option.
To solve this problem, an appropriate rule would prohibit the use of cross-entity ipso facto triggers as part of firm-wide enforcement but would permit them when they are part of project-specific enforcement. If the primary creditor is using cross-liability provisions in a way that pushes the entire firm into bankruptcy, then we should not allow it to turn around and pretend that it is acting against these entities individually. Prohibiting this particular use of the clauses still allows for selective enforcement but eliminates the use of tailored partitions to simply evade bankruptcy law.170
Such a rule need not be limited to instances where the primary creditor has turned to firm-wide action by choice. When the entities are sinking together in a way that the primary creditor cannot avoid, the cross guarantees are implicated, and firm-wide enforcement is the de facto path. The primary creditor’s option for selective enforcement exists only when one entity is still viable. Once both entities have filed for bankruptcy, the option has no value. Thus, the protection of the option is not in the enforcement of ipso facto clauses but rather in the court’s determination of good-faith filing.171 If viable entities are prohibited from filing bankruptcy, then the project-specific enforcement option is preserved, and none of the ipso facto clauses governing the viable entities’ obligations are subject to bankruptcy law and its prohibition on ipso facto clauses.
Judge Peck’s reading of the Bankruptcy Code in Charter Communications and Lehman Brothers provides for this outcome. He suggests that a clause that is triggered by the filing of bankruptcy by any affiliated debtor is a prohibited ipso facto clause. The reach of Judge Peck’s holding is only to cases where both affiliate entities file for bankruptcy. Because there is no state law prohibiting ipso facto clauses, any entity that has not filed for bankruptcy cannot rely on the Bankruptcy Code for protection. Thus, as I have suggested is optimal, only simultaneous bankruptcies172 will trigger the prohibition.173
IV. the selective enforcement theory: limitations and critiques
A. Creditor Opportunism
The greatest danger posed by selective enforcement is creditor opportunism. An unconditional cross guarantee of payment gives a primary creditor the option to enforce a covenant selectively against any one entity within a corporate group. The target entities need not even include the defaulting entity. This provides an ex post opportunity for primary creditors to use a hold-up threat to extract value.
Very little in the credit agreements protects against this. Some facilities include onerous covenants that put debtors in technical default very easily if a creditor wants to extort value. Relationships and repeat-play dynamics, however, play a strong role in preventing this situation. It is common for a primary creditor or creditor group to be headed by a bank with a long relationship and history with the debtor. This provides the creditors with useful information about the loan prospects, but it also provides the debtor with useful information about the likely behavior of its creditors. Banks have reputations and are repeat players with particular debtors and within industries. They also compete for debtors’ business. A creditor who has a reputation for being trigger-happy with defaults is not likely to be a favored candidate for a company shopping its primary debt. Nor are other creditors likely to favor that creditor’s involvement as a primary creditor.174 The likely result is that a debtor’s cost of capital will increase if it brings on a primary creditor with a reputation for opportunistic behavior.
If anything, the banks that tend to be primary lenders with the selective-enforcement option are a check on some general creditors (such as hedge funds) that swoop in when the company is in distress.175 Those funds have less need for a market reputation, since they can buy into the distressed debt market without the consent of other stakeholders. The primary creditors can use the selective-enforcement option to avoid broad enforcement actions that might cause the contagious spread of distress, which would trigger hold-up rights among those investors.176
In cases where the restraints on primary creditor opportunism are weak, the market has an additional tool at its disposal: covenant-light loans. Fears of lender opportunism may explain the increasing trend in recent years toward their use.177 Reducing the strength of covenants limits creditors’ ability to act opportunistically, but it also limits their valid enforcement options. Lenders competing to make a loan might reduce the number of covenants in an agreement to entice the debtor to borrow from them. It may be that debtors demand covenant-light loans in particular when they do not have a longstanding trust relationship with a creditor. The data on whether this is actually the case is murky.178
In the end, there is no bulletproof way to avoid creditor opportunism. But the dynamics at play—relationship lending, reputational considerations, covenant-light loans, and the like—all reduce the risk of opportunism posed by a primary creditor with a selective-enforcement option. Indeed, the primary creditor’s option to selectively enforce and thus limit the hold-up rights of other creditors may in some instances be a protection against opportunism from those creditors.
B. Differentiating Motives
This Article claims that the puzzle of the corporate web can be explained in many cases by an analysis of tailored partitions as a means to create selective-enforcement options. I do not claim, however, that all partitions and cross liabilities are attempts to create selective-enforcement options. As I note throughout, firms face all kinds of pressures to partition assets. Tax regulations, path dependency, accounting regulations, jurisdictional rules, and the like might also be driving partitions. But these explanations cannot account for the majority of large firms that go into bankruptcy with massive corporate webs divided by legal entities and connected by cross-liability provisions. Indeed, the puzzle for academics has been coming up with an explanation for these partitions that cannot be explained by jurisdiction, tax, or other regulatory factors. For example, one might speculate that a firm like Kodak partitioned its foreign subsidiaries for tax reasons or to comply with local regulations. But those were not the subsidiaries connected by cross guarantees that filed bankruptcy petitions together.179 Regardless of the foreign entities, we need an explanation for what was going on with the sixteen domestic entities that filed for bankruptcy in the Southern District of New York.
One challenge is that it may be difficult to distinguish between tailored partitions that are creating selective-enforcement options and those that exist for other reasons. Along the way, I have suggested several characteristics (for example, the presence of a primary creditor to all entities) that may be indicative. But the question merits future treatment. As the literature on entity partitioning matures, the ability to differentiate the various motivations behind those partitions will be invaluable to courts and scholars in determining the appropriate policy responses for a specific case.
I can say a little at this time about distinguishing some potential motives. Selective enforcement, withdrawal rights, and structural priority may each drive partitions. Structural priority is created whenever entities are partitioned.180 It is fairly easy to identify the effects of structural priority and to understand its impact on a claim. But it will rarely be clear whether structural priority is the purpose behind a partition or just a known incidental effect of it. Because the intent is hard to ascertain but the effects are easy to identify, structural priority should generally be respected because doing so does not affect the value of withdrawal rights or selective enforcement. And carefully written cross guarantees, subordination agreements, and other side agreements could actually eliminate structural priority if the relevant parties were determined to do so. There is therefore no obvious value in any rule that eliminates structural priority.
Withdrawal rights are more complicated. Withdrawal rights exist when a legal partition gives a creditor to one entity the ability to withdraw the assets in that entity from the entire enterprise upon default. That creditor can then effectively shut down a firm if the asset is critical. As I have discussed elsewhere, these rights create a powerful substitute for monitoring.181 This substitute has value especially when there is a risk of management misbehavior or major obstacles to monitoring a particular asset, or when a particular creditor is an ineffective monitor.
The key characteristics of withdrawal rights are that they run in one direction. The value does not exist when a creditor is the primary creditor on all relevant assets and those assets have been connected by cross guarantees. Moreover, when entities on both sides have withdrawal rights, the rights are more likely to be subject to abuse and to create costly confusion. Withdrawal rights should, therefore, be favored when they protect a peripheral creditor and run in one direction without bilateral cross guarantees.
In those cases where withdrawal rights that run in one direction are protecting a peripheral creditor, the separate entity should be treated like a third party. But this also means that courts can justifiably expect that it will have been run like one. The creditor with the withdrawal right may not be able to monitor performance, but it must monitor the separateness of the entities to ensure its withdrawal right is worth something and is visible to the world. The court then can expect that when a partition is created for withdrawal-rights purposes, assets and liabilities will not be commingled, and separate books will have been meticulously kept for the withdrawable entity. It should also expect the creditor of that entity not to be the primary creditor of the core entity of the enterprise.
In these ways, a withdrawal partition can be differentiated from a selective-enforcement partition. On a grander scale, this suggests that the treatments discussed above should be limited to cases in which the selective-enforcement option is at play. The presence of tailored partitions by way of overlapping cross-liability provisions will be key to this inquiry, as will the presence of a primary creditor, syndicate, or common group of creditors.
The law-and-economics literature of corporate groups generally focuses on the concept of seamless asset partitions.182 But firms today often divide partially related assets in a tailored fashion to create selective-enforcement options for a primary creditor. That primary creditor specializes in monitoring the enterprise as a whole. But when selective-enforcement options allow the creditor to precisely calibrate ex post enforcement, the benefits of that special expertise can be more fully realized. With enhanced enforcement capabilities, the creditor can reduce the risks to its investment from a debtor’s incompetence or misbehavior. Reduced investor risk, in turn, translates into a lower cost of capital for the borrowing firm.
On the other hand, selective enforcement may introduce costs through creditor opportunism. It is not clear, however, that such costs are significant. At the very least, the value of selective enforcement must be considered along with its costs when bankruptcy courts and policymakers grapple with complex cross liabilities and legal partitions. Indeed, a theory of selective enforcement provides insight into some of the most pressing questions in bankruptcy today. While the analysis above is preliminary and must be empirically tested, further analysis of fraudulent transfers, equity guarantees and stock pledges, good-faith filing, and the like should not ignore the important reality of the new corporate web.
Appendix: Specific Provisions
A. Cross Defaults/Cross Guarantees
A cross guarantee is a promise that one entity will pay the debts of another entity. A cross-default provision is merely a provision stating that a loan will go into default if the borrower defaults on another loan. When used across entities, a cross-default provision will cause the default of any debt of one entity to default on the major loan of another.
Here are two examples of the language that might be used in a cross guarantee:
Guaranty; (a) (i) Each of the Company and each US Subsidiary Guarantor, jointly and severally, hereby absolutely, unconditionally and irrevocably guarantees the punctual payment when due, whether at scheduled maturity or on any date of a required prepayment or by acceleration, demand or otherwise, of all obligations of each other Loan Party and each other Subsidiary of the Company now or hereafter existing under or in respect of the Loan Document . . . .183
Guarantee. Each Guarantor unconditionally guarantees, jointly with the other Guarantors and severally, as a primary obligor and not merely as a surety, the due and punctual payment and performance of the Obligations. Each Guarantor further agrees that the Obligations may be extended or renewed, in whole or in part, without notice to or further assent from it . . . .184
A cross-default provision may look like this:
[Defining default of the major loan to occur when any other]:
[D]efault shall occur under any Debt issued, assumed or guaranteed by any Borrower aggregating in excess of $250,000, or under any indenture, agreement or other instrument under which the same may be issued, and such default shall continue for a period of time sufficient to permit the acceleration of the maturity of any such Indebtedness for Borrowed Money (whether or not such maturity is in fact accelerated), or any such Indebtedness for Borrowed Money shall not be paid when due (whether by demand, lapse of time, acceleration or otherwise).185
[Defining default of the major loan to occur when any other]:
(e) Cross-Default. The Borrower or any Material Subsidiary (i) fails to make any payment when due (whether by scheduled maturity, required prepayment, acceleration, demand, or otherwise) in respect of any Material Indebtedness, or (ii) fails to observe or perform any other agreement or condition relating to any Material Indebtedness, or any other event occurs, the effect of which default or other event is to cause or to permit the holder or holders of such Material Indebtedness to cause, with the giving of notice if required, such Material Indebtedness to become due prior to its stated maturity; provided, however, that an Event of Default under this Section caused by the occurrence of a default with respect to such Material Indebtedness shall be cured for purposes of this Agreement upon the party asserting such default waiving such default or upon the Borrower or such Subsidiary curing such default prior to such party exercising any remedies with respect thereto if, at the time of such waiver or such cure the Administrative Agent has not exercised any rights or remedies with respect to an Event of Default under this Section;186
B. Cross Guarantees of Payment/Cross Guarantees of Collection
The baseline contract rule in most jurisdictions is that if the contract does not specify otherwise, lender has an absolute guarantee of payment and may enforce first against either A or B. This means that when A defaults, the lender can go after B for payment without ever going after A for any payment.
The parties may, however, contract around this rule. The contract may provide that the lender cannot call a default against B without trying to collect against A first. This allows a further tailoring of enforcement. The default rule is known as a guarantee of payment where the restricted guarantee is a guarantee of collection.
Not every jurisdiction has law on point. And in any event, the parties in the agreements we are discussing are sophisticated enough almost always to include precise language on which type of guarantee is being adopted. Because there is some uncertainty on the default rule, lending lawyers tend to include very precise language making it clear that the guarantee is an absolute guarantee of payment.
An example of such language might include the following:
This guaranty hereunder is a guaranty of payment and not of collection. Each Guarantor waives any right to require the Agent or any Lender to sue any Borrower or any other Guarantor, or any other Person obligated for all or any part of the Guaranteed Obligations (each, an “Obligated Party”), or otherwise to enforce its payment against any collateral securing all or any part of the Guaranteed Obligations.187
Guarantee of Payment. Each of the Guarantors further agrees that its guarantee hereunder constitutes a guarantee of payment when due and not of collection, and waives any right to require that any resort be had by the Administrative Agent or any other Secured Party to any security held for the payment of the Obligations or to any balance of any deposit account or credit on the books of the Administrative Agent or any other Secured Party in favor of any Borrower, any Account Party or any other Person.188