Before the crisis, securitisation was hailed as a major improvement to the functioning of financial markets. By allowing risk to be held more widely and banks’ balance sheets to be more liquid, securitisation had increased the supply of credit and was thought to have improved financial stability (Duffie 2007).
As a result, the issuance of asset-backed securities grew in leaps and bounds, peaking in 2006 at $4 trillion. However, the bursting of the US real estate bubble in 2006-2007 and the collapse of the subprime mortgage market, which precipitated the most severe financial crisis since 1929, indicated that securitisation also has a dark side. Implicated in the creation of a pyramid of opaque, complex, highly leveraged assets, the technique underwent a swift demise: in 2009-2010, securitised issuance slumped to about $1 billion a year, mostly from US government-sponsored enterprises and self-securitisations by European banks for refinancing activities with the ECB.
But does securitisation really deserve its unenviable reputation?
The basic issue with securitisation is information asymmetry. Banks make it their business to gather information on prospective borrowers so that they can grant and manage loans. But this information cannot be credibly transmitted to the market when loans are securitised. And, if it cannot be priced properly, banks could lack incentives to adequately screen borrowers at origination or to keep monitoring them once the assets are securitised.
These perverse incentives might be neutralised if banks could find ways to overcome or at least mitigate the presence of asymmetric information in the context of securitisation. For example, banks might choose to securitise loans that have a relatively low soft information content. Or they might retain a high share of the securitised portfolio’s risk by keeping the most junior (equity) tranche as a signalling device of its (unobservable) quality or as a device that commits the bank to keep monitoring borrowers (“skin in the game”). Further, as banks deal with investors on a repeated basis, reputational concerns may deter them from selling lemons – they might even choose to securitise loans of better-than-average quality to safeguard their reputation. In fact, similar dynamics have been observed for banks underwriting securities issued by firms that are also their borrowers (see Kroszner and Rajan 1994 for the 1920’s and Gande et al. 1997 for the 1990s).
The assessment of the impact of securitisation on the quality of lending standards is ultimately an empirical issue with important policy implications. The existing evidence on the whole supports the thesis that the rise of subprime mortgages was accompanied by a decline in lending standards (see Dell’Ariccia et al. 2008; Mian and Sufi 2009; Keys et al. 2010).
However, all these papers focus on aggregate data from the US subprime mortgage market, which is a small and unusual segment of the credit market as a whole (representing less than 10% of all securitised mortgages in the US and an insignificant fraction of the equivalent EU market). In particular, subprime mortgages are, by definition, much riskier than other mortgages and also involve a greater degree of asymmetric information, since they are granted to borrowers with little or no track record.1
New evidence from Italy
In a study of the larger part of the market for securitised assets, i.e. prime mortgages, in Italy, a country with a regulatory framework similar to those prevalent in Europe, we find results consistent with a very different behaviour by banks (Albertazzi et al. 2011). In particular, by analysing a unique dataset of about 1 million household mortgages originated by 50 Italian banks in the years 1995–2006, we find that:
- Securitised mortgages are on average less risky than mortgages with similar characteristics that banks keep on their books.
- When banks choose loans for securitisation, they take into account the need to reduce information asymmetry.
- Securitisation deals are also structured so as to reduce information asymmetry.
- These results and their dynamics across time can be explained by a commitment by banks to build and maintain the value of their reputation: This is consistent with banks’ traditional certification role observed in many markets (see Drucker and Puri 2009).
Based on these results we draw two main conclusions.
- First, securitisation does not necessarily affect lending standards. While it may do so in specific markets, such as the US subprime market, this by no means holds true in the more general case. In other words, the technique is not harmful per se.
- Second, fundamental underwriting practices matter, as do regulation, supervision and the institutional setting.
What are the policy implications?
First, prime mortgages can be securitised in a soundly functioning market, depending on factors such as the regulatory and supervisory framework. The lesson here is that loan originators should be kept close to their basic business. Supervisors should emphasise sound underwriting practices and good business practices irrespective of the ultimate securitisation structure. Just as rating a loan should involve expert judgment along with quantitative analysis, supervisory practices should rely on face-to-face interaction with banks as much as on regulatory ratios. Also moving in the right direction are the decisions of the Basel Committee on Banking Supervision that require that banks conduct more rigorous credit analyses of externally rated securitisation exposures and strengthen minimum capital requirements to better reflect the degree of complexity of securitisation assets and therefore their inherent risk (BCBS 2009).
Second, banks should always be aware that their reputation is at stake. Reputational mechanisms are likely to be most effective when investors can easily assess the behaviour of originating banks, which is more likely when the complexity of the assets is reduced to a minimum and appropriate information is available to investors. This implies standardisation and more transparency. A clear step in this direction is the Eurosystem’s ABS loan-level initiative for securities accepted as collateral in credit operations, which will introduce loan-by-loan information requirements for existing and newly issued ABS by mid-2012. The Bank of England’s initiative to require greater transparency in relation to structured products (ABS) as part of the eligibility criteria for instruments accepted in its operations should achieve similar results. Public and private initiatives are also underway in the US to enhance the quality and use of loan-level information – which has been available for some time but was rarely used by investors prior to the crisis.
The conclusion is that securitisation per se is not necessarily evil.
Disclaimer: The views expressed in this column are those of the authors and do not necessarily reflect those of their respective institutions.
Albertazzi, U, G Eramo, L Gambacorta and C Salleo (2011), “Securitization is not that evil after all”, Bank of Italy, Tema di discussione, no 741 and BIS Working Paper, no 341.
Agarwal, S, Y Chang and A Yavas (2010), “Adverse selection in mortgage securitization”, Paolo Baffi Centre Research Paper.
Basel Committee on Banking Supervision (2009): Enhancements to the Basel II framework, Basel, July.
Dell’Ariccia, G, D Igan and L Laeven (2008), “Credit booms and lending standards: evidence from the subprime mortgage market”, International Monetary Fund Working Paper, no 08/106.
Drucker, S and M Puri (2009), “On loan sales, loan contracting, and lending relationships”, Review of Financial Studies, 22(7),2835-2872.
Duffie, D (2007), “Innovations in credit risk transfer: implications for financial stability”, Stanford University Working Paper.
Elul, R (2009), “Securitization and mortgage default: reputation vs adverse selection”, Working Papers, Research Department of the Federal Reserve Bank of Philadelphia.
Gande, A, M Puri, A Saunders, and I Walter (1997), “Bank underwriting of debt securities: modern evidence”, Review of Financial Studies, 54:1175-1202.
Mian A and A Sufi (2009), “The consequences of mortgage credit expansion: evidence from the mortgage default crisis”, Quarterly Journal of Economics, 124(4), November.
Keys B, T Mukherjee, A Seru and V Vig (2010), “Did securitization lead to lax screening? Evidence from subprime loans 2001–2006”, Quarterly Journal of Economics, 125(1), January.
1 Evidence from the US prime mortgage market, based on partial loan-level datasets, is mixed (see, among others, Elul (2009) and Agarwal et al (2010)).