The transfer of credit risk can be described as a mode of risk management employed by financial institutions, often in the form of securitization or credit default swaps. Securitization has been a major theme of the post-crisis legislation in the EU and the US, partly due to the
originate-to-distribute model of lending and its causal role in the wave of defaults on
residential mortgages. The risk retention rule has in turn been adopted as a means of restoring sound credit-granting criteria and thus the incentive misalignment between originators and investors. Apart from its financial stability implications, the transfer of credit risk additionally alters the debtholders’ role in the field of debt governance, bringing about decision-making distortions, expressed in weakened monitoring and inefficient initiation of bankruptcy, escalating into the so-called empty creditor problem. Since the transfer of credit risk is the common denominator in both contexts, this paper argues that the risk retention rule might in turn have a positive impact on monitoring and decision-making upon insolvency, geared towards the preservation of firm value. Among the various options of risk retention, holding a vertical piece throughout the securitized exposure or a piece of the mezzanine tranches might approximate the original lender’s incentives more effectively, compared to equity tranches. Thus, the specific modality of slice held might carry divergent informational content about the banks’ stance towards screening and monitoring.
1. Introduction ... 1
2. The role of debtholders in corporate governance ... 2
2.1. Bank debt ... 2
2.2. Covenants and monitoring ... 4
3. Credit risk transfer instruments ... 6
3.1. Credit Default Swaps ... 6
3.2. Securitization ... 7
3.2.1. The mechanics ... 8
3.2.2. Incentives & the financial crisis ... 10
4. Post-crisis EU regulation: Risk retention ... 12
5. The transfer of credit risk in debt governance ... 14
5.1. Impact on covenants and monitoring ... 15
5.2. Impact on renegotiation and bankruptcy... 18
5.2.1. Credit Default Swaps ... 18
5.2.2. Securitizations ... 21
6. Counteractive forces & proposed solutions ... 23
6.1. Covenants and monitoring ... 24
6.2. Renegotiation and bankruptcy ... 25
6.3. A possible role for risk retention in debt governance ... 26
6.3.1. True sale securitizations ... 27
6.3.2. Synthetic securitizations ... 28
6.3.3. The choice of slice ... 29
7. Conclusion ... 32
Credit risk transfers are employed in the context of risk management by credit institutions. Nonetheless, the absence of credit risk can nudge the lender’s incentives towards inefficient behavior, removing the consequences of his actions. In particular, the transfer of credit risk can weaken the incentives to properly screen the borrower, before the loan origination, and monitor his performance, during the lending relationship.
Misaligned incentives are understood both on the level of the financial system, as well as in the narrower field of corporate (debt) governance.1 The first aspect (financial stability) has attracted major theoretical and legislative attention, owning to the magnitude of the recent financial crisis. In particular, the originate-to-distribute model2 has been held partly responsible for the wave of defaults, which further spurred extensive destabilizing effects, culminating into the recent financial crisis. On that premise, EU and US legislation have introduced the risk retention rule as a means of restoring the lenders’ skin-in-the-game and thus their original incentives, in order to prevent similar systemic shortcomings associated with securitizations.
Regardless of any financial stability considerations, since the absence of the lender’s exposure primarily alters the lending relationship as such, the transfer of credit risk concerns debt governance, as well. The detachment of control rights from credit risk might thus negatively affect monitoring, covenant design and waivers, as well as impact debt renegotiation and initiation of bankruptcy proceedings (the empty creditor problem). Weakened monitoring incentives and inefficient initiation of bankruptcy are in turn associated with a decrease in firm value, either by increasing the cost of capital or by destroying any going-concern surplus. Thus, it is important to note that the ballast is here shifted entirely from financial stability to decision-making inefficiencies within the narrow debtholder-firm relationship.
1 Debt governance is understood as the “creditors’ overall relationship with the debtor” including the “negotiations to address
loan terms and conditions”, as well as “the exercise or restructuring of contractual and legal rights”, Henry TC Hu and Bernard Black, ‘Debt, Equity and Hybrid Decoupling: Governance and Systemic Risk Implications’ (2008) 14 European Financial Management 663, 665, 681.
Keeping that distinction in mind, while risk retention was originally introduced on the background of systemic considerations, both contexts are commonly grounded on the function of the lending relationship, focusing either internally (debt governance) or externally (financial system). Since the cause of the problem -i.e. the transfer of credit risk- is common in both fields, the question that arises is whether risk retention can have positive spillovers in the context of debt governance, as well. Before drawing any conclusions, one should therefore first turn to the role of debtholders, especially banks, in corporate governance and then map the complications that arise due to the use of credit risk transfer instruments.
Utilizing the agency and informational asymmetries framework, one could recognize at least two functions of debt in curbing adverse incentives and/or resolving frictions, both purportedly linked to efficient outcomes.3 First, debt can objectively assume a disciplinary role in reigning the free cash-flow problem, ultimately pointing to the efficient allocation of capital.4 Second, from a subjective perspective, the lender is deemed to perform project selection and monitoring, which are also linked to allocative efficiency, mitigating informational asymmetries and moral hazard issues.
Project selection and monitoring are fundamental in the specific context of bank lending. As a matter of background, a bank’s primary function within the financial system is to channel savings (deposits) into the productive economy (loans), a process underpinned by the fractional
3 John Armour and others, Principles of Financial Regulation (OUP 2016) 53. Informational asymmetries and agency
problems are the main two frictions regarding the allocation of capital, Jeremy Stein, ‘Agency, Information and Corporate Investment’ in George M Constantinides, Milton Harris and René M Stulz (eds), Handbook of the Economics of Finance, vol 1 (Elsevier BV 2003) 114.
4 The free-cash flow problem indicates the managerial preference for size over profitability. Debt could induce discipline as
any failure to pay coupons is tied to default and its consequences, Michael Jensen, ‘Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers’ (1986) 76 The American Economic Review 323; Stein (n 3) 121.
reserve banking.5 Yet the particularity of the banking model does not lie in mobilizing capital, but rather in the bank’s expertise and cost advantage in project screening and monitoring.6
Screening refers to the scrutiny prior to the loan origination and is mostly guided by adverse
selection considerations, owning to asymmetric information.7 The latter left unchecked could either drive safe borrowers out of the market or induce lenders to ration credit.8 Monitoring, on the other hand, is performed during the lifecycle of the lending relationship and is driven by the lender’s motivation to recover his investment. In that sense, monitoring primarily targets moral hazard.9
It follows that on the other side of the lending relationship there are naturally small, non-traded firms, for whom market-based finance is by and large unfeasible, due to accentuated informational asymmetries.10 In the context of larger firms, the same banking function can add to the borrower’s reputation and thus indirectly rectify the investors’ information and coordination problems vis-à-vis the borrowing firm (delegated monitoring).11 Therefore, delegated monitoring has an alleviating effect on the informational asymmetry between the borrowing firm and the rest of the firm’s dispersed debtholders.12 As a consequence, bank debt can assume a certifying/signaling role, lowering the borrowing firms’ cost of capital.13
5 Armour and others (n 3) 275.
6 Allen Berger and Gregory Udell, ‘Relationship Lending and Lines of Credit in Small Firm Finance’ (1995) 68 The Journal
of Business 351, 354; Armour and others (n 3) 29, 287.
7 The asymmetric information/signaling framework is based on Akerlof’s and Spence’s seminal papers, George Akerlof, ‘The
Market for " Lemons ": Quality Uncertainty and the Market Mechanism’ (1970) 84 The Quarterly Journal of Economics 488; Michael Spence, ‘Job Market Signaling’ (1973) 87 The Quarterly Journal of Economics 355.
8 John Kiff, François-Louis Michaud and Janet Mitchell, ‘An Analytical Review of Credit Risk Transfer Instruments’ [2003]
Financial Stability Review 125, 110.
9 Douglas Diamond, ‘Financial Intermediation and Delegated Monitoring’ (1984) 51 The Review of Economic Studies 393.
Moral hazard refers to the incentive misalignment between parties in a contractual relationship, the most informed between them extracting value in unobservable ways, Jean-Jacques Laffont and David Martimort, The Theory of Incentives (Princeton 2002); Kiff, Michaud and Mitchell (n 8).
10 Berger and Udell (n 6) 354.
11 That is, it resolves the free-rider problem faced by multiple lenders or obviates multiple parallel monitoring efforts, Armour
and others (n 3) 29, 276.
12 Kiff, Michaud and Mitchell (n 8) 109.
13 Douglas Diamond, ‘Monitoring and Reputation : The Choice between Bank Loans and Directly Placed Debt’ (1991) 99
Journal of Political Economy 689; Alan Morrison, ‘Credit Derivatives, Disintermediation, and Investment Decisions’ (2005) 78 Journal of Business 621, 623.
In principle, if rational debtholders foresee the shareholders’ harmful behavior,14 they will adjust the terms of lending ex ante, effectively raising the firm’s cost of debt.15 Loan covenants, typically embedded in loan contracts, in turn restrict several concretely defined aspects of the borrowers’ financial condition and/or behavior that could harm the debtholders’ interests.16 In that sense, covenants limit future managerial discretion and thus lower the cost of debt, raising the value of the firm ex ante, while serving as a monitoring device.17
However, it should be noted that if covenants merely (re)distribute cash flows between claimholders, they should be irrelevant to the value of the firm, at least in principle.18 That said, loan covenants also limit investment decisions and are extensively employed in financial contracting, which is consistent with a possible role in resolving imperfections beyond the Modigliani/Miller setting, especially adverse selection and moral hazard issues.19
Apart from directly prescribing or prohibiting certain behavior, covenants form a subtler basis for monitoring via their acceleration clauses, which provide for the right to demand immediate repayment of the principal upon breach (i.e. before the original maturity).20 In that sense, debt maturity effectively becomes contingent on the lender’s scrutiny, who has to verify the breach
14 Including acts that just redistribute firm value or both redistribute and lower firm value (e.g. risk-shifting), Clifford Smith
and Jerold Warner, ‘On Financial Contracting. An Analysis of Bond Covenants’ (1979) 7 Journal of Financial Economics 117, 118.
15 ibid 119.
16 One could distinguish between affirmative and negative covenants, which prescribe or prohibit certain behavior and financial
covenants, which are based on financial ratios and are further subdivided into maintenance-based and occurrence-based, depending on the time/circumstances they need to be met; Greg Nini, David Smith and Amir Sufi, ‘Creditor Control Rights, Corporate Governance, and Firm Value’ (2012) 25 Review of Financial Studies 1713. Another typology distinguishes between covenants that restrict dividend policy, financing and investment decisions, as well as bonding covenants, Smith and Warner (n 14) 124.
17 To the extent that those benefits are set off against the costs of contracting (monitoring, bonding, enforcement) there is
arguably a firm-specific optimal set of financial contracts that maximize firm value, Ileen Malitz, ‘On Financial Contracting: The Determinants of Bond Covenants’ (1986) 15 Financial Management Association International 18, 19; Smith and Warner (n 14) 121.
18 The Irrelevance Hypothesis is based on the Modigliani/Miller Propositions, according to which capital structure just
repackages risk and return, those being fundamentally rooted on investment (as opposed to financing) decisions. In that sense, covenants have been paralleled to benign mutations, that neither harm nor benefit and thus persist over time, Malitz (n 17); Smith and Warner (n 14) 123.
19 Raghuram Rajan and Andrew Winton, ‘Covenants and Collateral as Incentives to Monitor’ (1995) 50 The Journal of Finance
1113; Smith and Warner (n 14) 154. See also n 7 and n 9.
and, by implication, to monitor the borrower. It follows that covenants function as “contractual devices that increase a lender’s incentive to monitor” (italics added).21
On the other hand, creditors can choose to waive the violation and renegotiate the existing contract terms. Thus, apart from delegated monitors, relationship lenders can be also recognized as “delegated renegotiators”.22 As a backdrop, if renegotiation fails, the lender retains the power to initiate formal bankruptcy proceedings.23 On that basis, lenders can exert significant informal influence over managerial decision-making, scaling up to the replacement of the existing management, with a view to maximize the possibility of recouping their investment.24
However, among the various modalities of covenants, the ones that pertain to investment decisions are particularly costly to monitor and enforce, sometimes even being virtually impossible to contract upon in the first place.25 To the extent that certain contingencies are unforeseeable, or even foreseeable but too many to contract or too costly to enforce,26 contracts are by definition incomplete.27 Nevertheless, since debt provides the debtholders with a certain, defined degree of control throughout the productive life of the firm (covenants)28 and completely cedes it upon bankruptcy, it is eventually conceptualized as contingent allocation of control. In that sense, debt determines a specific governance structure.29
Overall, to the extent that a subtler or more intrusive, yet informal, shift in control is thus entailed, covenants exemplify the lender’s inclusion in the broader scope of corporate governance,30 ultimately perfected by the definitive relocation of control to debtholders within
21 Rajan and Winton (n 19).
22 Marco Becht, Patrick Bolton and Ailsa Röell, ‘Corporate Governance and Control’ in George Constantinides, Milton Harris
and René Stulz (eds), Handbook of the Economics of Finance, vol 1 (Elsevier 2003) 29.
23 See n 20 and Smith and Warner (n 14) 152.
24 Nini, Smith and Sufi (n 16); Douglas G Baird and Robert K Rasmussen, ‘Anti-Bankruptcy’ (2010) 119 The Yale Law
Journal 648, 678.
25 Smith and Warner (n 14) 153; Michael Jensen and William Meckling, ‘Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure’ (1976) 3 Journal of Financial Economics 305, 336.
26 Jean Tirole, ‘Incomplete Contracts: Where Do We Stand?’ (1999) 67 Econometrica 741, 743–744.
27 Philippe Aghion and Patrick Bolton, ‘An Incomplete Contracts Approach to Financial Contracting’ (1992) 3 The Review of
Economic Studies 473, 490. Similarly Jaime Zender, ‘Optimal Financial Instruments’ (1991) 46 The Journal of Finance 1645.
28 Aghion and Bolton (n 27) 492.
29 The other two varieties of control are entrepreneur and investor control (in the Aghion and Bolton model), ibid 490–491. 30 Consistently, “[…] a corporate governance problem arises whenever an outside investor wishes to exercise control
bankruptcy proceedings.31 What latently binds this whole narrative together, though, is that the lenders’ incentives ultimately stem from their own exposure to the cash flow variability of the firm, i.e. to the (credit) risk assumed. By implication, it can be safely inferred that any change in the credit risk borne, especially a complete transfer, could alter the debtholders’ incentives, all else being equal. However, the change in incentives is sensitive to the specific credit risk transfer instrument used, among which credit default swaps and securitizations are of particular interest to debt governance.
Credit default swaps (CDSs) are derivatives, whose payoff depends on the creditworthiness of a specified entity (reference entity).32 The party seeking to insulate itself from the risk of default, e.g. a lender (protection buyer), contracts with a protection seller, who agrees to pay the face value of the debt (notional principal)33 issued by a reference entity, upon the latter’s default (credit event). The protection seller, in exchange, receives a premium, typically in periodic payments.34
As instruments functionally equivalent to insurance against credit risk,35 credit default swaps are typically employed as part of financial institutions’ risk management, banks being the commonest users among them.36 Nonetheless, unlike other means of risk management, there is no transfer of title. The immediate effect is that the lender’s position, while legally identical,
ways in which suppliers of finance to corporations assure themselves of getting a return on their investment”, Andrei Shleifer and Robert Vishny, ‘A Survey of Corporate Governance’ (1997) 52 The Journal of Finance 737.
31 Reinier Kraakman and others, The Anatomy of Corporate Law (OUP 2017) 109. After all, bankruptcy is not synonymous to
liquidation, but could just as well end up in reorganization, Aghion and Bolton (n 27) 490.
32 John Hull, Options, Futures and Other Derivatives (Pearson 2015) 573.
33 That is accurate for physical settlement; in case of cash settlement, the protection seller pays the difference between the face
value and the amount recovered at the ISDA organized auction, ibid.
34 Jonathan Berk and Peter DeMarzo, Corporate Finance (Pearson 2020) 785. The amount paid periodically, expressed as a
percentage of the notional principal, is the CDS spread. To the extent that CDSs protect against the risk of default, a corporate bond yield netted against the CDS spread should approximate the risk-free rate, Hull (n 32) 573–575.
35 The term ‘insurance’ is used here in a loose, rather than a technical sense, Hull (n 32) 574. 36 ibid 571.
loses its economic underpinning. To the extent that its exposure is now insured, the bank in effect no longer bears the credit risk and thus the consequences of its further decisions. Crucially, though, credit default swaps differ from insurance in that there is no insurable interest required.37 As a result, credit default swaps can serve speculative rather than hedging purposes,38 which accounts for both the multiplicatively larger size of the credit protection market over the debt market, as well as for the implied incentives.
The term securitization describes a technique by which relatively illiquid assets, such as consumer or residential loans, are transformed into liquid ones, namely tradable securities.39 Securitization results in the issuance of asset-backed securities (ABSs) which derive their payoff from an underlying loan portfolio (pooling) in a sequenced manner (tranching), bearing any loan default-related losses in reverse order (subordination).
While typically obscure in its mechanics, securitization ultimately unburdens the loan originator from a relatively illiquid loan portfolio’s credit risk, which is ultimately borne by investors.40 As a baseline remark, pooling implies that securitization’s rationale rests primarily with diversification and the entailed neutralization of idiosyncratic risk. Furthermore, while
tranching and subordination just repackage the risk and return, they also imply that another
rationale for securitization lies with the scaled informational sensitivity of the tranches, which in turn improves liquidity.41 However, as it will be analyzed below, liquidity often comes at the cost of efficient decision-making, especially regarding screening and monitoring.42
37 ibid 574.
38 Berk and DeMarzo (n 34) 785.
39 Günter Franke and Jan Pieter Krahnen, ‘The Future of Securitization’ (2008) 31 8; Christian Farruggio and André Uhde,
‘Determinants of Loan Securitization in European Banking’ (2015) 56 Journal of Banking and Finance 12, 13.
40 Hull (n 32) 185; Franke and Krahnen (n 39) 8.
41 Peter DeMarzo, ‘The Pooling and Tranching of Securities : A Model of Informed Intermediation’ (2005) 18 The Review of
Financial Studies 1, 2; Günter Franke, Markus Herrmann and Thomas Weber, ‘Loss Allocation in Securitization Transactions’ (2012) 47 The Journal of Financial and Quantitative Analysis 1125, 1126.
42 Christine Parlour and Guillaume Plantin, ‘Loan Sales and Relationship Banking’ (2008) 63 The Journal of Finance 1291,
Securitization first includes pooling a certain number of loans or a range of other debt instruments, which are then sold to a special purpose vehicle (SPV), thus taken off the originator’s balance sheet (typically a bank). Already at this point, the originator has theoretically shed the portfolio-related credit risk, has improved liquidity43 and has freed up lending capacity, having realized the proceeds from the sale and because of the now improved capital adequacy, all else being equal.44 Further down the same process, the SPV finances the purchase by issuing asset-backed securities (ABSs) and secondarily asset-backed commercial paper (ABCP).45 ABSs’ payoff in turn derives primarily from the underlying loan cash flows,46 so that their issuance can be thought of as finalizing the transformation of the original loan portfolio.
On a technical level, the cash flows from the loans (underlying assets) are divided and assigned to different, sequenced classes of securities (tranching), which receive payments in a waterfall-like manner47 and thus bear losses in reverse order (subordination). To a bare minimum, there are three of them, namely the senior, the mezzanine and the equity tranche.48 A second step taken, the same process can be repeated for mezzanine tranches, whose cash flows are repackaged anew, underpinning the issuance of another sequence of securities, based on preexisting ABSs (ABS CDOs).49
Typically, each subsequent tranche carries lower rating and higher promised returns, owning to the payment order, the relative size and characteristics (e.g. maturity and risk) of the loan
43 Securitization as a means of funding can be a stand-alone motivation, Farruggio and Uhde (n 39); Clara Cardone-Riportella,
Reyes Samaniego-Medina and Antonio Trujillo-Ponce, ‘What Drives Bank Securitisation? The Spanish Experience’ (2010) 34 Journal of Banking and Finance 2639.
44 Steven Schwarcz, ‘The Future of Securitization’ (2009) 41 Connecticut Law Review 1313.
45 As a matter of typology, mortgage-backed securities (MBSs) are based on residential mortgage loans. If the underlying
assets include a wider range of debt instruments, the securities are termed collateralized debt obligations (CDOs), Hull (n 32) 583.
46 Though indirectly, through the SPV. The issuance of ABCP serves as a supporting mechanism, being used for ABS payments
in case of irregular cash flows from the underlying loans, Armour and others (n 3) 439.
47 That is, every subsequent tranche receives principal and coupon payments only if the former has been fully paid first. There
are two waterfalls, one for the principal and one for the interest payments, Hull (n 32) 187.
48 ibid 186.
49 ibid 188. Another form of resecuritization are the so-called CDOs squared, whose underlying assets are other CDOs, which
in turn might be based on MBSs, Kathryn Judge, ‘Fragmentation Nodes: A Study in Financial Innovation, Complexity, and Systemic Risk’ (2012) 64 Stanford Law Review 657, 682.
portfolio portion assigned to it. As a result, tranches also differ as to their informational sensitivity. Thus, unlike junior tranches, senior tranches largely capture the benefits of diversification, so that their value should be relatively insensitive to firm-specific information.50 Nonetheless, the different risk properties among tranches are by no means static. Instead, it has been shown that changes in the underlying portfolio, like an increase in the probability of default or correlation, may shift the inter-tranche (relative) loss distribution in favor of junior tranches.51
The transfer of credit risk by means of true sale can nevertheless be replicated by retaining the loan portfolio ownership and buying protection against default, by entering into a credit default swap contract with investors (synthetic securitization).52 Crucially, though, unlike true sale transactions, synthetic securitizations are also consistent with speculative purposes, since the same set of assets can be referenced multiple times.53 Additionally, since synthetic securitizations do not involve the transfer of the underlying assets, control rights remain with the original lender. As a result, true sale and synthetic securitizations might have similar results on the investor side,54 but diverge significantly as far as the originator and his incentives are concerned.
It should be noted that the main three levels implied above, namely the originator, the special purpose vehicle and the investor level, do not strictly correspond to the actual parties involved, which are usually more (outsourcing). For example, in mortgage lending, the originator typically concludes the loan contract, possibly via a mortgage broker, while the funds are provided by another entity, the warehouse lender. The lending relationship is then transposed to the special purpose vehicle level and managed by the servicer, who collects the loan payments and handles defaults. At the same time, the underwriter, typically an investment bank, handles the issuance of the asset-backed securities, which are rated by a credit rating agency. The party most involved throughout the process, the arranger, sets up the special
50 Franke and Krahnen (n 39) 13.
51 Jan Pieter Krahnen and Christian Wilde, ‘Risk Transfer with CDOs’ (2008) 15 12; Franke and Krahnen (n 39) 36. 52 Franke, Herrmann and Weber (n 41) 1127.
53 Angelos Delivorias, ‘Understanding Securitisation’ (2015) 15. 54 ibid 7.
purpose vehicle and concludes the web of contractual relationships among the parties that constitute the securitization conduit.55
Therefore, securitizations reshape lending relationships both by untying the credit risk from the lender’s balance sheet, as well as by transforming bilateral relationships into multi-party structures. While both elements affect the incentives involved, the former has attracted major attention following the recent financial crisis.
As a matter of framework, it can be said that there is a trade-off between optimal risk allocation, akin to market finance, and efficient decision-making, which is akin to bank finance.56 However, pre-crisis securitizations, ambitiously combining some elements of both types of finance, did poorly in both respects, combining inefficient decision-making and poor risk allocation.57
In particular, securitization turned out Janus-faced, facilitating risk management at the expense of sound credit-granting criteria, which devolved into the origination of poorly assessed loans, only to be distributed through securitization schemes (originate-to-distribute model). Put differently, originators had little incentive to scrutinize potential lenders, as the credit risk would be borne by investors (moral hazard).58 The resulting expansion of lending activity created a self-feeding loop of credit expansion and asset appreciation, that scaled up to the creation of the real estate bubble, setting the stage for the recent financial crisis.59 Additionally, the complex structure of mortgage securitizations and the inherent difficulty to assess the underlying assets, combined with the lack of appropriate rating methodology60 hindered the understanding of risk assumed by investors and could be further associated with mispricing
55 Franke and Krahnen (n 39) 17–18. 56 ibid 10; Farruggio and Uhde (n 39). 57 Franke and Krahnen (n 39) 10.
58 Markus Brunnermeier, ‘Deciphering the Liquidity and Credit Crunch 2007–2008 Banking Industry Trends Leading Up to
the Liquidity Squeeze’ (2009) 23 Journal of Economic Perspectives 77, 82; Armour and others (n 3) 414.
59 While a price increase would normally lower demand, the opposite trend can be explained by the residential real estate being
both a consumption good and an investment asset. Investment assets appreciate on expectations (anticipated capital gains), which at times significantly depart from fundamentals (asset bubble), Richard Dusansky and Çagatay Koç, ‘The Capital Gains Effect in the Demand for Housing’ (2007) 61 Journal of Urban Economics 287; Berk and DeMarzo (n 34).
(adverse selection).61 Furthermore, the higher returns of asset-backed securities, compared to equally rated bonds, positively stimulated demand on the investor side, exacerbating the bubble.62
The inevitable defaults on residential mortgages eventually spurred the reverse downward spiral in asset prices. The following uncertainty about the value of asset-backed securities resulted in the widespread refusal to roll over asset-backed commercial paper, a situation equivalent to a bank run.63 The underlying maturity mismatch, although seated at the special purpose vehicle level,64 spread its consequences to the originating banks, through the credit lines (liquidity puts) embedded on the securitization vehicles,65 which were nonetheless imperfectly accounted for in capital requirements under Basel I.66 Eventually, the originating banks had their own solvency exposed, having to meet relatively sudden and massive liquidity needs (funding risk).67
Thus, contrary to its perceived risk reduction properties, securitization failed to allocate risk outside the banking sector and facilitated increased risk taking by originators, eventually being conducive to individual as well as systemic vulnerability.68
Among the various weaknesses of securitization, poor originating incentives, though mediated through multiple channels, have thus been a partial cause for the pronounced effects of structured finance on financial stability. Post-crisis legislation has taken those effects into
61 Rational ignorance may also be involved, Judge (n 49) 692; Cem Demiroglu and Christo James, ‘The Dodd–Frank Act and
the Regulation of Risk Retention in Mortgage-Backed Securities’ in Paul Schultz (ed), Perspectives on Dodd-Frank and Finance (MIT Press 2015) 206; Hull (n 32) 194.
62 Hull (n 32) 194; Judge (n 49) 679, 695. 63 Brunnermeier (n 58) 94; Judge (n 49) 700.
64 Since ABCP is typically short-term and overcollateralized, it has been thought of as equivalent to a deposit, which
nonetheless financed long-term assets, Armour and others (n 3) 440; Brunnermeier (n 58).
65 The originator providing liquidity puts can be seen as a private lender of last resort, nevertheless operating without the
potency of a central bank, Armour and others (n 3) 440.
66 Wenying Jiangli and Matt Pritsker, ‘The Impacts of Securitization on US Bank Holding Companies’ [2008] SSRN Electronic
Journal 1, 4; Brunnermeier (n 58); Armour and others (n 3) 440; David Jones, ‘Emerging Problems with the Basel Capital Accord: Regulatory Capital Arbitrage and Related Issues’ (2000) 24 Journal of Banking and Finance 35.
67 Rob Nijskens and Wolf Wagner, ‘Credit Risk Transfer Activities and Systemic Risk: How Banks Became Less Risky
Individually but Posed Greater Risks to the Financial System at the Same Time’ (2011) 35 Journal of Banking and Finance 1391, 1392; Brunnermeier (n 58) 80.
68 Nijskens and Wagner (n 67); Jiangli and Pritsker (n 66); Francesca Battaglia and Angela Gallo, ‘Securitization and Systemic
Risk: An Empirical Investigation on Italian Banks over the Financial Crisis’ (2013) 30 International Review of Financial Analysis 274, 275.
account in mandating risk retention on the originator or sponsor level, primarily targeting the originate-to-distribute model and the resulting incentive misalignment between originators and investors (moral hazard).69
Despite the multi-dimensional shortcomings of securitization, the EU policy attempts to remedy the underlying weaknesses, instead of rejecting securitization altogether. Thus, post-crisis EU legislation, specifically the Securitization Regulation,70 attempts to re-establish securitization in the context of Capital Markets Union.71
Among the new rules, the specific problem of incentive misalignment between originators and investors is targeted through the risk retention rule, which has been also introduced in the US by the Dodd-Frank Act in 2010.72 According to art. 6(1) of the EU Securitization Regulation, the originator, the sponsor73 or the original lender shall keep a material net economic interest in the securitization of at least 5%, on an ongoing basis.74 This mainly translates into either a vertical slice of 5% in each tranche, a horizontal slice in the first-loss piece (equity tranche) or a slice of the mezzanine tranches, always amounting to 5% of the total exposure. Importantly, the retained exposure cannot be subject to any risk mitigation or hedging.75
The immediate rationale of the rule is that risk retention partly reconstructs the incentives that induce properly assessed loan origination, that is, reflects a skin-in-the-game approach.76 That approach is analogous to the Jensen and Meckling agency framework, in that the originator’s (agent) incentives converge to the investors’ (principals) interests, as the fraction of the cost
69 Demiroglu and James (n 61) 205.
70 Regulation (EU) 2017/2402 of the European Parliament and of the Council [2017] OJ L347. 71 See p. 2 of the preamble to the Securitization Regulation, ibid.
72 Demiroglu and James (n 61) 201.
73 The originator according to art. 2 of the Securitization Regulation (n 70) is the entity that entered into the agreement that
created the exposure or acquired the exposure subject to securitization, while the sponsor is the entity that issues the securities that conclude the securitization.
74 Preamble of the Securitization Regulation (n 70), p. 10;Jan Pieter Krahnen and Christian Wilde, ‘Skin-in-the-Game in ABS
Transactions: A Critical Review of Policy Options’ (2018) 549.
75 Art. 6(1) Securitization Regulation (n 70). 76 Demiroglu and James (n 61) 201.
borne by the former increases.77 The difference is that the agent’s payoff in our case derives from the screening effort, rather than the extraction of perquisites;78 similarly, the cost implied by exposure drives the effort, instead of discouraging the appropriation of private benefits. In that sense, risk retention aims to address implications stemming from the separation between ownership and control, in the particular form found in securitizations (moral hazard).79 Nonetheless, given the different risk characteristics of the tranches, the flat-rate requirement of 5% has been subject to criticism.80 Additionally, confined only to originators/sponsors, risk retention has been criticized as a narrow and thus inadequate response against the complex agency problems posed by the securitization chain.81
Risk retention can be additionally analyzed in the framework of informational asymmetries.82 As a matter of background, buyers-investors, aware of their informational disadvantage and the seller’s incentive to overstate the value of the assets (adverse selection), will rationally lower the price they are willing to offer, eventually driving sellers of higher-than-average products out of the market. In a less stylized framework, though, it is more plausible that investors will ultimately bear the adverse selection costs.83 In that setting, risk retention can be conceptualized as a signal, credibly conveying the quality of the underlying assets and mitigating the problem of adverse selection.84 Nonetheless, the signaling value of risk retention can be doubtful, as it is now mandated, rather than chosen;85 however, the signaling rationale could remain for any percentage of the exposure held in excess.
77 Jensen and Meckling (n 25); Ingo Fender and Janet Mitchell, ‘Incentives and Tranche Retention in Securitisation : A
Screening Model’ (2009) 289 13; Krahnen and Wilde (n 51) 13.
78 Fender and Mitchell (n 77) 13.
79 Yingjin Hila Gan and Christofer Mayer, ‘Agency Conflicts, Asset Substitution and Securitization’ (2006) 12359 5. 80 Ethan Mobley, ‘Regulating Moral Hazard: The True Risk of Dodd-Frank’s Risk Retention Requirement’ (2017) 10 Business,
Entrepreneurship & the Law 45, 58; Demiroglu and James (n 61) 202.
81 Amy McIntire, ‘Dodd-Frank’s Risk Retention Requirement: The Incentive Problem’ (2014) 33 Banking & Financial
Services Report 11.
82 See n 7.
83 Guixia Guo and Ho Mou Wu, ‘A Study on Risk Retention Regulation in Asset Securitization Process’ (2014) 45 Journal of
Banking and Finance 61.
84 Hayne Leland and David Pyle, ‘Informational Asymmetries, Financial Structure, and Financial Intermediation’ (1977) 32
The Journal of Finance 371; DeMarzo (n 41); Craig Furfine, ‘The Impact of Risk Retention Regulation on the Underwriting of Securitized Mortgages’ [2018] SSRN Electronic Journal 1.
In any case, risk retention, though explicitly targeted to the incentive misalignment between originators and investors, targets the common cause of several, heterogeneous problems that also touch upon debt governance. Since the transfer of credit risk bears on almost every aspect of the lending relationship, it follows that risk retention is also expected to have an unintended positive effect on those problems as well, either directly or indirectly. Before turning to that question, one should therefore first map the several debt governance complications that follow the transfer of credit risk.
A recurring theme in the preceding analysis is the discrepancy between decision-making and consequences, especially in the background of financial stability. As analyzed above, the transfer of credit risk removes the conditions necessary for efficient decision-making, weakening the incentives to screen and monitor. By implication, the basic banking function (expertise in screening and monitoring) is altered, bringing about negative welfare effects. Crucially, the latter placed in a heavily interconnected financial system with its own structural weaknesses might have pronounced destabilizing effects that eventually spread out to the real economy.86
However, the discrepancy between decision-making and its consequences has a parallel expression in the context of debt governance. In that case, it is conceptualized as the divergence between the control rights and the exposure of the lender.87
While that description loosely fits both credit default swaps and securitizations, an important distinction is in order. Unlike true sale securitizations, credit insurance and synthetic securitizations involve the retention of title over the loans, nonetheless decoupled from the actual exposure. It follows that the original debtholder retains the same legal position and thus control rights within the lending relationship, but have effectively shed the exposure that underpins it - hence the term empty creditor.88 What further exacerbates the incentive problem,
86 Brunnermeier (n 58) 91–98. 87 Hu and Black (n 1) 680. 88 ibid.
though, is that the holder of control rights often has interests that oppose those of the borrowing firm, due to credit insurance, a situation that often implies speculative purposes.
By contrast, in true sale securitizations the title is transferred to the special purpose vehicle (issuer), the exposures are securitized and borne by investors, while the relationship is onwards managed by the servicer/trustee or by the originating bank. It follows that the problem in that case ultimately lies with the disassembly of the lending relationship, which brings about an analogous decoupling between control rights and cash flow rights.89 However, the main problem in that case appears to be rational apathy more so than moral hazard or speculative incentives.
Nonetheless, regardless of the particular form of decoupling, the alignment between control rights and exposure is an assumption that underpins the conceptualization of debt as contingent allocation of control, framed by the debtholders’ contractual rights along the productive life of the firm, as well as their statutory rights in bankruptcy proceedings (control rights).90 The control conceded to debtholders is in turn justified by the alignment of interests implied by exposure, in the sense that maximizing the value of debt claims concurs with maximizing the value of the firm, as well.91 Thus, the discrepancy between control rights and their economic substance, whatever its form, can bear on various stages of the lending relationship, from its inception (contract design) to its end (bankruptcy), often bringing about inefficient results.92
In the context of debt governance, the transfer of credit risk might alter the incentives to monitor, as well as its contractual pillar, i.e. covenants.93 In particular, the protected lender is
89 Demiroglu and James (n 61) 201; Florian Gamper, ‘Credit Default Swaps and the Empty Creditor Hypothesis - If It Ain’t
Broke Don’t Fix It’ (2015) 9 Journal of Business, Entrepreneurship and the Law 681.
90 Hu and Black (n 1) 681; Baird and Rasmussen (n 24) 678.
91 Henry Hu, ‘Corporate Distress, Credit Default Swaps, and Defaults: Information and Traditional, Contingent, and Empty
Creditors’ (2018) 32 Brooklyn Journal of Corporate, Financial & Commercial Law 5; Edward Janger and Adam Levitin, ‘One Dollar, One Vote: Mark-to-Market Governance in Bankruptcy’ (2019) 104 Iowa Law Review 1857.
92 Hu (n 91) 18.
93 Frank Partnoy and David Skeel, ‘The Promise and Perils of Credit Derivatives’ (2007) 75 University of Cincinnati Law
Review 1019, 1033; Hu and Black (n 1) 685. For credit protection, see Charles Whitehead, ‘The Evolution of Debt: Covenants, the Credit Market, and Corporate Governance’ (2009) 34 The Journal of Corporation Law 641. For securitization see Gary Gorton and George Pennacchi, ‘Banks and Loan Sales Marketing Nonmarketable Assets’ (1995) 35 Journal of Monetary
expected to rest upon the protection offered by credit default swaps, loosening the strictness of performance-based covenants, as well as collateral requirements.94 After all, if monitoring is costly, a bank protected against credit risk would rationally abstain from dedicating resources that do not increase its final payoff.95 Similarly, banks that frequently resort to securitizations are found to impose less restrictive interest coverage, debt-to-cash-flow and book leverage ratio requirements in covenants.96 This is consistent to the hypothesis that the mere access to the securitization market suffices to weaken the monitoring incentives, mostly expressed through performance-based covenants,97 even for non-securitized loans.98
That set of dynamics might in turn negate the positive effects induced by the lender’s scrutiny on the borrower decision-making, which is especially concerning when it comes to investment decisions.99 Left unchecked, the borrower of securitizing banks could engage in increased risk-taking,100 possibly scaling up to negative present value investments. Framed in broader terms, enhanced liquidity may come at the cost of ex ante inefficiency.101 Similarly, credit default swaps, associated with both reduced monitoring and increased probability of financial distress,102 might prompt risk-shifting behavior by the borrower, especially in case of over-insurance.103 As a matter of background assumptions, the over-protected lender might be especially incentivized to drive his borrower into bankruptcy, in order to capture rents from over-insurance, even if continuation is feasible.104 It is then hypothesized that the borrower Economics 389; Parlour and Plantin (n 42); Yihui Wang and Han Xia, ‘Do Lenders Still Monitor When They Can Securitize Loans?’ (2014) 27 Review of Financial Studies 2354, 2366.
94 Susan Chenyu Shan, Dragon Yongjun Tang and Andrew Winton, ‘Market versus Contracting: Credit Default Swaps and
Creditor Protection in Loans’ [2015] SSRN Electronic Journal; Chenyu Shan, Dragon Yongjun Tang and Andrew Winton, ‘Do Banks Still Monitor When There Is a Market for Credit Protection?’ (2019) 68 Journal of Accounting and Economics 101241.
95 René Stulz, ‘Credit Default Swaps and the Credit Crisis’ (2009) 15384 7. 96 Wang and Xia (n 93) 2355.
97 Shan, Tang and Winton, ‘Do Banks Still Monitor When There Is a Market for Credit Protection?’ (n 94) 7. 98 Parlour and Plantin (n 42) 1292.
99 Wang and Xia (n 93) 2360.
100 However, weakened monitoring and increased risk are not inefficient as such, i.e. if not translated into negative npv
investments, ibid 2371, 2387.
101 Parlour and Plantin (n 42) 1294. 102 See below, section 5.2.1.
103 Chanatip Kitwiwattanachai and Jiyoon Lee, ‘The Effect of Credit Default Swaps on Risk Shifting’ [2014] SSRN Electronic
Journal.
104 Even if the projects are of positive net present value, untimely bankruptcy is possible for liquidity reasons, Patrick Bolton
and Martin Oehmke, ‘Credit Default Swaps and the Empty Creditor Problem’ (2011) 24 Review of Financial Studies 2617, 2618.
would in turn better-off taking higher risks ex ante, lowering the probability of his projects succeeding, rather than risking his lender over-insuring.105
On a larger scale, loosened incentives to monitor are at odds with what distinguishes banks from other intermediaries in mobilizing capital, that is, their superior capability to monitor. To the extent that the agency problems that bank debt mitigates are now reintroduced, the cost of capital is expected to increase,106 implying not only the reshuffling of risk and return,107 but also the expected inefficient behavior by the borrowing firm. As a result, firm value might decrease ex ante.108 Similarly, the firm’s cost of capital might increase in the bond market, as the signaling value of bank debt vis-à-vis dispersed and less informed third investors could be doubtful.109 Consistently, the increase in bond spreads associated with credit protection is expected in case of riskier and informationally opaque firms,110 for whom bank debt is especially beneficial in terms of processing ‘soft’ information, mostly unobservable but important for the firm’s future performance.111
Nonetheless, the interaction between poor investment decisions, the likelihood of default and credit default spreads can be a counteractive factor to this shortcoming. In particular, it is hypothesized that lenient covenants will be less frequent for borrowers prone to agency problems, given the higher cost to buy protection against them.112 Nevertheless, this argument does not apply to securitizations, where the loosening of covenants, as well as the subsequent increased risk taken on by the borrowing firm have no equivalent market-based counteractive force.
Furthermore, as mentioned earlier, creditors typically utilize acceleration clauses in the covenant design. Covenant violations are nevertheless often waived, but the lender in turn
105 Murillo Campello and Rafael Matta, ‘Credit Default Swaps and Risk-Shifting’ (2012) 117 Economics Letters 639. 106 Adam Ashcraft and João Santos, ‘Has the CDS Market Lowered the Cost of Corporate Debt?’ (2009) 56 Journal of
Monetary Economics 514, 523. On the other hand, the ability to hedge risks can in turn lower the bond spreads. The increased informational capacity of CDSs can have the same effect, ibid 515. In securitizations, the cost of debt might be similarly decreased due to lower liquidity premiums, Parlour and Plantin (n 42) 1293.
107 See n 100.
108 Ivo Welch, Corporate Finance (4th edn, 2017) 529.
109 Amar Gande and Anthony Saunders, ‘Are Banks Still Special When There Is a Secondary Market for Loans?’ (2012) 67
Journal of Finance 1649; Morrison (n 13); Ashcraft and Santos (n 106).
110 Ashcraft and Santos (n 106) 515. 111 Parlour and Plantin (n 42) 1293–1294.
112 Shan, Tang and Winton, ‘Market versus Contracting: Credit Default Swaps and Creditor Protection in Loans’ (n 94) 4;
imposes tighter restrictions, which often pertain to investment and financing decisions113 and might even scale up to direct control on the firm’s management.114 This function is in turn altered in case of credit risk transfers, absent its own foundation, namely the lender’s exposure. Banks actively engaging in securitizations are indeed found to frequently waive covenant violations without imposing additional restrictions on the terms of the loan contract, such as increased interest rates or additional collateral requirements.115
The mapping of incentives so far relies on the absence of credit risk, which is translated into the lack of incentives typically attributed to lenders. As such, it would apply mutatis mutandis to both credit default swap protection and securitizations. In a more nuanced approach, though, credit default swaps and (true sale) securitizations differ regarding the mode of transferring risk, which is reflected on incentives as well. In particular, credit default swaps involve a positive payoff upon default due to insurance, combined with the retention of control rights, while securitization results in the fragmentation of an otherwise bilateral lending relationship, due to the transfer of title. 116 Thus, while both forms of risk transfer are consistent with a passive approach to the lending relationship, only credit insurance is further associated with the active engagement in inefficient behavior, motivated by the direct CDS payment upon default.
Approaching insolvency, it is hypothesized that the lender will be less motivated to compromise in debt renegotiations, resting upon the CDS payoff upon default,117 even when restructuring would be otherwise optimal.118 However, it is argued that the borrower, aware of his future reduced bargaining position would in turn be disciplined ax ante, since he would be
113 For example, prescribing the reduction of investments or leverage, Sudheer Chava and Michael Roberts, ‘How Does
Financing Impact Investment? The Role of Debt Covenants’ (2008) 63 Journal of Finance 2085; Michael Roberts and Amir Sufi, ‘Control Rights and Capital Structure: An Empirical Investigation’ (2009) 64 Journal of Finance 1657.
114 Baird and Rasmussen (n 24) 678. 115 Wang and Xia (n 93) 2357.
116 Gamper (n 89); Baird and Rasmussen (n 24).
117 Shan, Tang and Winton, ‘Market versus Contracting: Credit Default Swaps and Creditor Protection in Loans’ (n 94) 2;
Shan, Tang and Winton, ‘Do Banks Still Monitor When There Is a Market for Credit Protection?’ (n 94) 2.
better-off avoiding defaults altogether.119 Looser covenants for referenced firms are also arguably consistent with that mode of market-induced discipline.120 In that line of thought, credit default swaps are characterized as an ex ante commitment device that reduces strategic defaults.121 Nonetheless, the borrower’s discipline would not change the lender’s lack of motivation to engage in private workouts, even when it would be efficient to do so.122 In that sense, CDSs might result in increased credit risk and likelihood of financial distress, owing to lenders’ reduced renegotiation motivation,123 and thus bring about negative effects on firm value and investment.124
Furthermore, since credit protection dissociates the lenders’ interests from the value of borrowing firms, yet without prejudice to control rights, lenders can be conceptualized as empty
creditors.125 In that sense, the lack of incentives to engage in out-of-court restructurings might be further transformed into the positive motivation to induce bankruptcy proceedings, since the latter provides a certain, full payoff.126 Protected lenders are thus better-off opting for bankruptcy, even if an out-of-court restructuring would be efficient preserving any going-concern surplus, which is now destroyed. After all, unlike the uncertain restructuring outcome, the protected lenders’ payoff is certain in case of default127 and, crucially, unaffected by the deadweight costs normally incurred in bankruptcy.128 In fact, lower recovery means higher (direct) CDS payment.129 In a similar sense, a principal write-down or a debt-for-equity swap might be preferable over the liquidation value, but crucially not over the CDS payment, an
119 Bolton and Oehmke (n 104); Stefan Arping, ‘Credit Protection and Lending Relationships’ (2014) 10 Journal of Financial
Stability 7.
120 Shan, Tang and Winton, ‘Do Banks Still Monitor When There Is a Market for Credit Protection?’ (n 94) 5, 22. 121 Bolton and Oehmke (n 104).
122 ibid; Daniel Hemel, ‘Empty Creditors and Debt Exchanges’ (2010) 27 Yale Journal on Regulation 159.
123 Marti Subrahmanyam, Dragon Yongjun Tang and Sarah QianWang, ‘Does the Tail Wag the Dog?: The Effect of Credit
Default Swaps on Credit Risk’ (2014) 27 Review of Financial Studies 2926.
124 Stefano Colonnello, Matthias Efing and Francesca Zucchi, ‘Empty Creditors and Strong Shareholders: The Real Effects of
Credit Risk Trading. October 2016 Draft.’ [2016] SSRN Electronic Journal.
125 Hu and Black (n 1) 682; Baird and Rasmussen (n 24) 680–681.
126 This outcome though depends on the definition of the credit event, which might also include restructuring. Yet voluntary
debt exchanges do not typically qualify as credit events (see section 4.7, 2003 ISDA Credit Derivatives Definitions); Stulz (n 95) 8; Hemel (n 122) 163.
127 Stulz (n 95) 9.
128 Bolton and Oehmke (n 104) 2632.
129 In the sense that CDS contracts typically stipulate that the seller shall pay the difference between the recovery and the
assessment that is finally expressed in the preference for the realization of the credit event, regardless of the entailed value destruction.130
At the extreme, over-insured debtholders might even have a strong, positive interest in leading the borrowing firm into bankruptcy, in order to reap the higher CDS payment, even if that outcome destroys value.131 In those circumstances, over-insured CDS holders have in fact negative net exposure to the firm, in that their interests are inverse to the interests of the borrowing firm, in what has been described as negative economic ownership.132 After all, over-insurance always offers a payoff higher than any possible renegotiation surplus and might be even conceptualized as a bet against the reference firm.133 Worse still, since credit protection might just as well serve speculative purposes, a protection buyer might acquire the reference entity’s debt only afterwards, with the sole purpose of holding out restructuring, induce bankruptcy and thus realize on his net short position.134
In the same context, a similar but more nuanced approach factors in the arguably positive welfare effects of credit protection, prompted by the borrower’s ex ante discipline analyzed above. In that setting, the latter is expected to raise the surplus conceded in renegotiation, in order to nudge the lender into renegotiations, while committing to refrain from (costly) strategic defaults. The rationale is that the lender, who has strengthened his bargaining position due to credit protection, will be now offered a larger share of the restructuring surplus value, which would in turn prevent unnecessary bankruptcy filings. Even so, in the presence of over-insurance, the lender would in any case pursue (inefficient) liquidation, to the extent that his payoff from the CDS contract is larger than any renegotiation surplus possibly offered.135
130 Hu and Black (n 1); Hemel (n 122); William Bratton and Adam Levitin, ‘The New Bond Workouts’ (2018) 166 University
of Pennsylvania Law Review 1597. Nonetheless, the empirical evidence is mixed, see David Mengle, ‘The Empty Creditor Hypothesis’ (2009) 3; Mascia Bedendo, Lara Cathcart and Lina El-Jahel, ‘Distressed Debt Restructuring in the Presence of Credit Default Swaps’ (2016) 48 Journal of Money, Credit and Banking 165; András Danis, ‘Do Empty Creditors Matter?’ (2013) 34/2013.
131 Partnoy and Skeel (n 93) 1035; Baird and Rasmussen (n 24) 681. Interestingly also characterized as Darth Vader monitors,
Partnoy and Skeel (n 94) 1035.
132 Or net short debt activism, Hu (n 91); Hu and Black (n 1); Hemel (n 122); Janger and Levitin (n 91). 133 Hemel (n 122) 168.
134 Stulz (n 95) 9; Hemel (n 122) 167; Janger and Levitin (n 91) 1865; Bratton and Levitin (n 130) 1635. For cases, see Henry
Hu, ‘Financial Innovation and Governance Mechanisms: The Evolution of Decoupling and Transparency’ (2015) 70 The Business Lawyer 347, 369; Hu (n 91) 21.
Synthetic securitizations present parallel possibilities of decoupling control rights and cash flow rights, which in turn underlie the same empty creditor problem, possibly also escalating into a positive interest to see the value of referenced firms decrease. In that case, the originator remains the owner of the assets, which then become the reference portfolio in credit default swap contracts with investors. As a result, the holders of securitization positions bear the credit risk associated with a pool of underlying assets, though by means of credit default swaps rather than asset-backed securities.136 In that sense, the originator retains the control rights, though separated from the credit risk of the assets in the reference portfolio, similarly to a lender that has acquired protection. At the extreme, the originator might employ the same asset in multiple synthetic securitizations, which effectively shifts his interests into negative ownership as in over-insurance, in the sense described above.137
By contrast, the issue of loan renegotiation in true sale securitizations is shaped on a different basis, as the lending relationship is handled by the servicer (mortgage-backed securities) or the indenture trustee (collateralized debt obligations).138 Thus, initial screening and monitoring, as well as loan renegotiation and filing for bankruptcy can be performed by different parties, though it is not uncommon for the originator to also perform servicing.139 Nonetheless, unlike origination, servicing implies a continued tie to the lending relationship until its resolution, in or out of bankruptcy.
Handling the investors’ interests rather than his own, the servicer/trustee can be thus described as an agent vis-à-vis the holders of securitization positions (principals).140 However, bondholders lack the ability or even the interest to contract on the terms of servicing, being typically dispersed and of heterogeneous interests, while often underestimating the servicing risk.141 As a result, there is an agency complication unattended by contracting.
136 ‘The EBA Report on Synthetic Securitization’ (2015) 7. 137 Hu and Black (n 1) 688.
138 Janger and Levitin (n 91) 1868; Hu and Black (n 1) 686; Gan and Mayer (n 79) 2.
139 Vinod Kothari, Securitization. The Financial Instrument of the Future (Wiley 2006) 673, 696.
140 Steven Schwarcz, ‘Keynote Address: The Conflicted Trustee Dilemma’ (2010) 54 NYLS Law Review 707, 708; Adam
Levitin and Tara Twomey, ‘Mortgage Servicing’ (2011) 28 Yale Journal on Regulation.
141 Tomasz Piskorski, Amit Seru and Vikrant Vig, ‘Securitization and Distressed Loan Renegotiation: Evidence from the
Subprime Mortgage Crisis’ (2010) 97 Journal of Financial Economics 369, 370; Judge (n 49) 684, 702–703; Schwarcz (n 140) 708; Levitin and Twomey (n 140) 6.
Thus, assuming that the servicer is a third party, his incentives to renegotiate and engage in private workouts, if any, can be partly inferred by the mode of his compensation. It is important to note that the servicer is typically compensated on the basis of a flat-rate fee, dependent on the outstanding principal, rather than loan performance.142 Additionally, he is typically reimbursed for any costs related to foreclosure, but not renegotiation.143 The impact on renegotiation is nevertheless ambivalent. On the one hand, the prospect of continued servicing fees might nudge the servicer to delay bankruptcy or to unnecessarily modify the loan contract.144 On the other hand, speeding up foreclosure is also likely as a “low-cost exit”, to the extent that foreclosure costs are typically reimbursed, unlike the costs of renegotiation (foreclosure bias).145 In any case, though, the servicer’s incentives are disconnected from maximizing the value of the loan pool.146
Thus, in broader terms, delegating the relationship management to a third party effectively creates another type of empty “creditor”,147 who exercises control rights without any meaningful personal stake, that is, without internalizing the costs and benefits of his decisions.148 Indeed, it has been observed that servicers of mortgage securitizations were inclined to ensure automated and cost-minimizing processes, yet often uncoordinated and without any regard to the actual recovery.149 At the same time, servicing agreements often provided poor guidance and left little margin of discretion as to loan modifications. Additionally, since the agent serves divergent economic interests related to each tranche, his stance towards loan modifications is necessarily ambivalent.150
142 John Geanakoplos, ‘Solving the Present Crisis and Managing the Leverage Cycle’ [2010] FRBNY Economic Policy Review
101, 119; Gan and Mayer (n 79) 2.
143 Piskorski, Seru and Vig (n 141) 370.
144 Franke and Krahnen (n 39) 18; Adam Ashcraft and Til Schuermann, ‘Understanding the Securitization of Subprime
Mortgage Credit’ (2008) 9.
145 Sumit Agarwal and others, ‘The Role of Securitization in Mortgage Renegotiation’ (2011) 102 Journal of Financial
Economics 559; Piskorski, Seru and Vig (n 141) 370; Geanakoplos (n 142) 119.
146 Levitin and Twomey (n 140) 5. 147 Hu and Black (n 1) 687.
148 Armour and others (n 3) 444; Levitin and Twomey (n 140) 5. However, it is reported that servicers do sometimes hold the
first loss piece, Gan and Mayer (n 79) 3.
149 Armour and others (n 3) 444; ‘Problems in Mortgage Servicing from Modification to Foreclosure’ (2010) 1 353–354;
Levitin and Twomey (n 140) 4.
Therefore, the servicer/trustee’s limited involvement, both in terms of role and economic stake, combined with the multitude of interests involved (investors) and the multitude of borrowers within the pool, is consistent with rational apathy towards loan modifications.151 The latter explains not only the bias towards foreclosure but also the inefficiency of the modification as such, when chosen.152
On the other hand, since inter-party coordination within the securitization scheme is based on contract,153 the question that follows is whether the contractual definition of each party’s role in the process is enough to prevent or amend incentive distortions. The standard contracting device, damages, while sensible at first sight, might nonetheless prove inadequate in terms of tying a party’s actions to its actual economic interests. After all, the parties typically enjoy limited liability and might as well be undercapitalized, while the composite nature of the transaction further perplexes proving fault.154 Furthermore, since monitoring extends over the entire duration of the loans, contractual incompleteness imply that not all future contingencies can be captured beforehand.155
The analysis above implies that the problems caused by the transfer of credit risk vary considerably, depending on the mode of the transfer and the aspect of the relationship they refer to. In that sense, existing literature has employed divergent approaches, often presenting focused proposals relating to the specific manifestations of the problem, utilizing different theoretical backgrounds. Nonetheless, since the unifying theme remains the transfer of credit risk, it could be argued that risk retention can be also considered as an equivalent mechanism that essentially affects the common cause behind the wider range of debt governance implications. Before discussing the extent of that positive effect, one should take into account
151 Hu and Black (n 1) 687; Schwarcz (n 140) 709; Levitin and Twomey (n 140) 5. 152 Agarwal and others (n 145) 575.
153 Fender and Mitchell (n 77). 154 Franke and Krahnen (n 39) 18. 155 Schwarcz (n 140) 708.
the existing proposals, which can be subdivided into market-oriented approaches and proposals of concrete legal intervention.
The first approach implies that market forces will counteract several aspects of the problems identified above. In this line of thought, it is argued that the sensitivity of credit derivatives to new information about credit risk, incorporated into credit spreads, could be a market-based alternative to covenant-induced discipline.156 Indeed, safe and transparent referenced firms have their cost of capital decreased in the bond market,157 which is consistent to the argument that credit default swaps can substitute for the information-producing role of banks,158 at least for a subset of firms. Nonetheless, a generalization of that argument would be feasible if credit derivatives covered a wider range of firms.159 Thus, this approach essentially circumscribes the problem instead of presenting a way to mitigate it.160
A similar argument is that if a bank is a repeat player in the loan sales market, reputational concerns might prompt corresponding monitoring incentives, even after a credit risk transfer has taken place.161
Another, more direct but less realistic proposal is the establishment of a market, where lenders and protection sellers would trade on control rights. The rationale is that such a market would eventually close the gap between those that own the control rights, but lack the corresponding incentives -on the supply side- and those who are incentivized to exercise them, on the demand side.162 While the protection seller’s involvement is a sensible solution, it appears that a more feasible channel would be a contractual one, namely the protection seller demanding information by the protection buyer-lender, on the basis of the credit default swap contract.163
156 Whitehead (n 93) 128–129. 157 Ashcraft and Santos (n 106) 515. 158 Bolton and Oehmke (n 104) 2624. 159 Whitehead (n 93) 119, 130. 160 ibid 130.
161 ibid 125. For hold-out behavior in restructuring, see Hemel (n 122) 166.
162 Yesha Yadav, ‘The Case for a Market in Debt Governance’ (2014) 67 Vanderbilt Law Review 771.
163 Partnoy and Skeel (n 93) 1033. Although, in general, it is unusual for marker participants to deviate from the standard
