RBA Annual Conference – 2007 Discussion

Jonathan Fiechter

The turmoil in financial markets in recent weeks has demonstrated the value of conferences like this. They provide a chance to take stock of trends and emerging practices in our financial systems and to exchange views on the implications of these practices for financial stability.

Nigel Jenkinson and his two co-authors have produced a timely and thoughtful paper. As shown in Nigel's presentation, financial innovations, including structured finance products and financial derivatives, have led to a rapid growth in financial assets over the past 15 years, far in excess of the growth rate of the real economy.

A primary question for this session, and an issue likely to be debated in many capitals around the world in coming months, is whether financial innovation and the world of ‘unfettered finance’, to borrow Martin Wolf's description, contribute to financial stability. Credit risk that once might have been concentrated locally is now sliced and diced and distributed broadly across the globe. Is this a good thing?

In my brief comments, I want to focus on the effect of structured finance, and in particular, the securitisation of residential mortgages, on financial stability. I want to expand on the discussion of the effects of asset securitisation – what is referred to in Nigel's paper as ‘arms-length financing’. My comments may be equally applicable to other types of credit but given the time constraints, I want to focus on residential mortgage finance.

I agree fully with Nigel's conclusions that financial innovations and structured finance have transformed the credit granting process. I also agree that financial innovation has the potential to improve the overall performance of our financial system.

But I take a more cautionary stance in assessing to what extent financial innovation has provided substantial benefits to the overall economy. My proposition is that securitisation and the development of structured finance have resulted in an increase in the overall level of risk in the financial system.

Let me be clear, upfront, that a higher level of risk in the financial system is not necessarily a bad outcome, so long as it is properly understood, valued and priced.

I readily accept the many benefits that arise from being able to structure credits in a way that allows them to be spread across a broad investor base. But I believe we are still learning the extent to which the structured finance process has transformed the nature of the credit intermediation function.

I would like to make three observations to amplify these points.

My first observation is that the growing complexity of financial instruments and financing structures, such as collateralised debt obligations (CDOs), has exceeded the ability of many of us to understand the underlying risks. Many of these instruments and structures are not well understood in the market, have complex features that are difficult to model and, until recently, had not been tested in an environment of tight liquidity and a material level of defaults of the underlying assets.

I doubt that the ultimate investors in the various tranches of many of the new structured finance products fully understand the performance characteristics and default probabilities of the assets backing up the securities.

To illustrate this point, I would like to share with you the experience of a friend of mine who has been in the mortgage business for decades as a mortgage aggregator. He purchases pools of residential mortgages, subject to the individual loans meeting certain quality benchmarks. He has a team of analysts and with the aid of an expensive software program, the team analyses the default probability and loss potential of each mortgage loan based on factors such as the neighbourhood in which the property is located, the health of the local economy and the financial characteristics of the borrowers. In the past, he might typically have rejected about 5 per cent of the mortgage loans in the pool. Over a year ago, he reported that his rejection rate had risen to around 40–45 per cent. As a result, he dropped out of the securitisation business. When he notified the bankers who were selling him the mortgages of his decision, they told him it was not a problem – that it was easy to sell these loans to other securitisers.

This highlights a fundamental question regarding structured products. Is it really practical for an investor in a CDO, which may include various tranches of mortgage-backed securities, to go through the same type of analysis of the underlying mortgage loans that my friend went through? Rather than having whole loans to analyse, the CDO may be comprised of the riskier tranches of the mortgage-backed securities. Given the difficulty of analysing or placing a value on a CDO, it is likely that many investors in CDOs end up placing significant reliance on the credit rating of the CDO and the name and reputation of the entity that has set up the CDO.

A key question then is whether there are steps that could be taken to improve the transparency of CDOs and methodologies to facilitate more accurate valuations.

My second observation is that when the underlying mortgage assets default, loan workouts under these CDO structures will be far more difficult than for a portfolio lender or under a plain vanilla mortgage-backed security structure. As a result, models which rely on the historical default and loss rates of residential mortgages held in portfolio may underestimate the losses that may arise for mortgage loans that have been securitised.

When there is a general downturn in the economy, a banker (or mortgage insurer) will immediately contact borrowers who miss one or more monthly payments, and attempt to restructure the loans with the objective of keeping the borrowers in their homes. There is a well-accepted axiom in the lending business that the first loss is the best loss. In fact, pro-active bankers will identify ‘at-risk borrowers’ and contact them in advance of default to restructure the loans. Portfolio lenders in the US are quietly doing this right now for residential mortgage borrowers they deem as high-risk.

But for loans that have been packaged under these more complex financing structures, there are many more parties of interest. It may be more difficult to get the pre-approval of investors in the various tranches (each with a different default exposure) to permit the loan servicer to enter into negotiations with an at-risk borrower who has not yet defaulted on a loan. And not surprisingly, loan servicers, with limited credit exposure and no ongoing relationship with the borrower (unlike a portfolio lender), may be less aggressive than portfolio lenders in pursuing problem borrowers.

As a result, it may be that the historical probabilities of default and losses given default of mortgage loans held by portfolio lenders are not applicable to mortgage loans that have been securitised. Under the new financing model, the incentives (and ease) of working with troubled borrowers may no longer be the same.

My third observation relates to a statement in the paper that ‘… primitive risk does not disappear through financial engineering’. For a given level of credit risk in the financial system, spreading the risk ‘a mile wide and inch deep’ has obvious benefits. Because the credit risk is no longer concentrated in one or more lenders, the default of a pool of mortgage loans or credit cards or the failure of a large corporation no longer poses the same risk to individual institutions. Instead, the risk is spread across a number of investors in small bites.

But because the pain of a default is spread so widely, there has been an observable increase in the willingness of lenders to extend credit to higher-risk individuals and corporations so long as they are able to transfer some or all of the credit risk. If a mortgage broker, who earns fees based on the volume of loans originated, is several steps removed from an investor, then that broker may be more willing to extend credit further out on the risk curve to higher-risk borrowers. Similarly, a lending officer will have an easier time getting loans approved by the bank's credit committee when the loans are being originated for sale rather than for the bank's balance sheet.

I would posit that the result has been an increase in the overall level of credit risk being underwritten in the financial markets as borrowers that in the past might not have qualified for bank credit were granted loans. It is beneficial to society and to economic growth to broaden the range of borrowers with access to formal financial credit. And of course, there is nothing wrong with the provision of credit to weak borrowers, be they corporations or individuals, so long as both parties understand the risk and the credit is priced accordingly.

But the securitisation of financial credit may also have led to a loosening of underwriting standards. The combination of high levels of investable funds (easy money) and the insulation of the underwriting process from the assumption of credit risk appears to have resulted in reduced underwriting discipline. This has enabled the production of a high volume of loans that carry a high risk of default.

Mortgage loans that include features such as no income verification for the borrower (so called ‘no-doc’ or ‘liar’ loans), 100+ per cent financing (often in the form of ‘piggyback’ loans that combine a first and second loan), starting rates of interest at below-market rates (‘teaser rate’ loans), and monthly payments that do not cover interest owed (negative amortisation loans) are not new. Portfolio lenders have offered loans that contained one or two of these features to select borrowers such as high net worth individuals or professional real estate investors for years. There have been few, if any, problems.

But in the past several years, the use of these instruments has exploded in response to a willingness on the part of the capital markets to purchase these ‘affordability’ mortgage loans even when issued more broadly to the general public. And more problematic, these mortgage loan features have been layered on top of each other. More curiously, the capital markets assigned a relatively small risk premium to securities comprised of these ‘affordability’ loans even when they were made to borrowers with tarnished credit histories – the so-called sub-prime category. Until recently, the demand from the capital markets for mortgage products appeared insatiable.

As a result, in the past several years, borrowers have been given access to loan amounts well in excess of what they might have qualified for in years past. While such loans are too risky for most portfolio lenders, the concern appears to have been reduced for loans that were originated, securitised and sold to a third party. A question for all of us from a public policy perspective is: what are the consequences when underwriting standards for loans originated for sale and distribution are less rigorous than those applied by lenders for their own portfolio? Is there a relevant role for government to step into?

This question is well illustrated by the very popular 2/28 and 3/27 loans. These are mortgage loans for which the initial interest rate is fixed for two or three years at a below-market rate of interest. After the initial two- or three-year grace period, the interest rate is reset (by as much as 400 to 600 basis points) to the fully indexed rate. Some borrowers taking out these loans were qualified on the basis of the initial teaser rate rather than the fully indexed rate. An implicit assumption of such a loan was that the borrowers would refinance before it resets to the higher rate – hence, they were sometimes referred to as ‘bridge’ loans.

All of this points to a potential vulnerability in the ‘originate and distribute’ model – an apparent reduction in credit underwriting standards. Interestingly, the disruption appears to have been greatest in the money markets of the developed countries. As markets re-price some of the riskier securities, it will be interesting to observe any changes that are demanded by investors and lenders related to the level of transparency of structured products, the retention of risk by loan originators, the equity contribution by the sponsor, etc.