The future of securitisation: how to align
This article reviews the recent collapse of global securitisation markets and the loss of investor confidence in them. It then sets out measures that could be taken to revive and strengthen the securitisation process, including mechanisms based on retention requirements for originators. It ends with a number of simple implications for policymakers and market practitioners.
JEL classification: G100, G200.
Large losses in the value of mortgage-related products and an associated deterioration in investor appetite led to broad-based distress in securitisation markets from the summer of 2007. Problems started with subprime mortgagerelated instruments, which experienced severe credit quality deterioration as a long period of appreciating house prices in the United States came to a halt. Losses were magnified by increasingly illiquid markets, and worsened further during the broad investor retreat from risk triggered by the Lehman Brothers bankruptcy and subsequent signs of global recession. 2
As a result, spreads on securitised products soared (Graph 1, left-hand panel) and activity across most market segments came to a sudden stop. Issuance volumes, which had risen to a combined annual total for the United States and Europe of about $3.8 trillion over the 2005–07 period, collapsed to just over $2 trillion in 2008. Reflecting a generalised loss of investor confidence, most of this remaining issuance was in the US agency sector (ie securities underwritten by US government-sponsored mortgage financing enterprises) and in European securitisations used for refinancing activities with the ECB. The US subprime and Alt-A market, which had peaked at some $815 billion in 2006, vanished, as did markets for many other securitised instruments (Graph 1, right-hand panel).
The views expressed in this article are those of the authors and do not necessarily reflect those of the BIS or the National Bank of Belgium. Any errors and omissions remain those of the authors, who thank Emir Emiray for assistance with the data and graphs.
See Chapter II of the BIS 79th Annual Report (2009) for a five-stage description of the crisis.
BIS Quarterly Review, September 2009
AAA tranche spreads1
ABS issuance volume5
US subprime (lhs)
United States (rhs)
00 01 02 03 04 05 06 07 08
In basis points. 2 Spreads on CDS index contracts referencing AAA-rated tranches of US commercial mortgage-backed securities (CMBX index, series 3). 3 Ten-year student loan ABS spreads to one-month Libor. 4 Ten-year floating credit card loan spreads to one-month Libor. 5 In billions of US dollars; includes agency and private label securitisations.
Sources: JPMorgan Chase; SIFMA; UBS; BIS calculations.
Problems in the securitisation process were central to this collapse in activity. Securitisation involves the pooling of assets and the subsequent sale to investors of claims on the cash flows backed by these asset pools. 3 As such, securitisation tends to incorporate a rather long chain of participants and its functioning depends crucially on whether the relationships between these participants preserve discipline and maintain adequate information flows along that chain.
This article sets out measures that could be taken to revive and strengthen the securitisation process, thereby revitalising the flow of credit to sectors such as consumer and mortgage finance. Renewing securitisation has conjunctural as well as structural elements. Chief among the former is the large overhang of securitised products (ie the so-called legacy assets) sitting on bank balance sheets and the uncertainty regarding future asset performance, both of which are depressing valuations. In order to help markets recover, governments in a variety of countries have taken steps to remove this overhang – either in the form of “bad bank” and similar measures targeting bank balance sheets directly 4 or by reviving investor interest through the provision of government funding in the markets for particular securitisations. One example is the US Term Asset-Backed Securities Loan Facility (TALF), which provides loans on a collateralised, non-recourse basis to holders of certain types of newly issued asset-backed securities.
In the remainder of this special feature, the term securitisation will be used both for “traditional” asset-backed securities (ABS) backed by large homogeneous asset pools, such as credit card and auto loans, and for collateralised debt obligations (CDOs) and related instruments, which are backed by smaller pools of more heterogeneous assets. In addition, it will be assumed that the liabilities backing these asset pools are tranched, forming a threetiered capital structure of equity/first-loss, mezzanine and senior tranches.
See Fender and Scheicher (2009) for a rationalisation of these measures based on evidence of sizeable illiquidity premia in prices for certain subprime mortgage securitisations.
BIS Quarterly Review, September 2009
Overall, these measures may be showing some signs of success. Spreads on securitised products have come down from their peaks (Graph 1, left-hand panel) and volumes have recovered somewhat, though unevenly across market segments. However, while providing temporary relief, these measures are unlikely to attract the stable base of dedicated longer-term investors needed for securitisation markets to recuperate in a sustained fashion. With large parts of the traditional investor community (such as structured investment vehicles (SIVs) and other conduits) having disappeared, more needs to be done. Key to rebuilding investor confidence is addressing the structural weaknesses in securitisation that have been exposed by the crisis. These, and proposals to eliminate them, are reviewed below, focusing in particular on plans for originators and arrangers to retain some exposure to the securitisations they help to generate. The key finding is that the degree to which the originator’s retained stake will be affected by a downturn will significantly influence the impact that the stake will have on incentives to adequately screen borrowers.
The remainder of this article is organised as follows. The next section briefly describes securitisation markets and how they work. This is followed by sections focusing on structural shortcomings revealed by the crisis and ways to address them. The last section concludes by identifying some implications for policymakers and market practitioners alike, including a set of simple “rules” for the design of tranche retention schemes.
Tackling the structural weaknesses in securitisation
Securitisation: a short review of the basics
(i) What is securitisation?
… is the tranching
The starting point for any discussion of structural weaknesses in securitisation markets is the securitisation process. In general, securitised instruments can be defined through three distinct characteristics: (1) pooling of assets (either cash-based or synthetically created); (2) delinking of the credit risk of the collateral asset pool from that of the originator, usually through the transfer of the underlying assets to a finite-lived, standalone special purpose vehicle (SPV); and (3) tranching of liabilities (ie issuance of claims with different levels of seniority) that are backed by the asset pool. 5
A key aspect of tranching is the ability to create one or more classes of securities accommodating different investor appetites. One way to achieve this is to generate some tranches whose rating is higher than the average rating of the underlying asset pool (other tranches, in turn, will carry lower ratings or remain unrated) or to generate rated securities from a pool of unrated assets. This is accomplished through the use of various forms of credit support to create securities with different levels of seniority. The main tool in this context is the priority ordering of tranches with regard to the allocation of losses
See Fender and Mitchell (2005) for a broader discussion of these issues.
BIS Quarterly Review, September 2009
(ie subordination): the equity or “first-loss” tranche absorbs initial losses up to the level where it is depleted, followed by mezzanine tranches which take some additional losses, again followed by more senior tranches. As a result, the most senior claims are expected to be insulated – except in particularly adverse circumstances – from the default risk of the asset pool through the absorption of losses by subordinated claims.
Another type of credit support is provided through structural provisions based on triggers and threshold levels. One example is overcollateralisation tests, which, when triggered, divert cash flow to senior note holders, in an attempt to maintain stability of performance for these tranches over time. Another example is rules regarding the use of excess spread, which represents the difference between the income earned on the asset pool and contracted payments to the tranched liabilities. Excess spread tends to be accumulated for the benefit of all investors, but is released to equity holders once certain requirements are met.
In principle, these structural provisions can be used interchangeably with subordination. For example, a reduction in the credit support provided to senior tranches via subordination can be compensated through more stringent rules for releasing accumulated excess spread to equity tranche holders. A downside of these trade-offs is additional complexity and the associated analytical burden for investors: the evaluation of a securitised instrument (ie a tranche) cannot be confined to estimating the loss distribution of the asset pool alone. It is also necessary to model the distribution of cash flows from the asset pool to the tranches under different scenarios, based on an assessment of subordination and the deal’s structural features (CGFS (2005)).
… as a way to
(ii) Market organisation and incentives
One implication of the pooling and tranching that characterises securitisation markets is the need to involve a relatively large number of parties in the securitisation process (Graph 2 illustrates the range of participants for a generic transaction). Organising such a process in ways that maintain incentives (eg in terms of screening asset quality) and the flow of information along the chain of participants can be a challenge. For certain types of securitisations, this is now universally recognised to have gone wrong in the run-up to the current crisis.
The process starts with the originators, who extend loans or other forms of credit to ultimate borrowers. Those originators who, in t...