The impact of the subprime-crisis on european banks

Table of Contents

List of Figures

List of Tables

List of Abbreviations

1 Introduction
1.1 Motivation and Objective
1.2 Course of the Investigation

2 Fundamentals of the Subprime-Crisis
2.1 The US Housing and Subprime Mortgage Market
2.2 Characteristics of Subprime Mortgages
2.3 Business Model of US Mortgage Brokers
2.4 Financial Instruments Underlying the Subprime-Crisis
2.5 Consequences of the Fragmented Securitization Process

3 The Development of the Subprime-Crisis
3.1 Situation of the US Housing Market up to 2007
3.2 Timeline of the Subprime-Crisis in 2007
3.3 Spillover Effects from the Mortgage Market to the Global Capital Markets
3.4 Consequences for the British Banking Market

4 Empirical Analysis About the Subprime-Crisis
4.1 History and Overview of Event Studies
4.2 Framework of an Event Study
4.3 Selection of Relevant Data
4.3.1 British Banks and Market Index
4.3.2 News about Private Financial Institutions and Central Banks
4.4 Event Study About the Subprime-Crisis
4.4.1 Event Study Methodology
4.4.2 Formulation and Testing of Hypotheses
4.4.3 Interpretation of Results
4.5 Year-round Performance of the British Banking Sector in 2007

5 Summary and Conclusion

References

Appendixes

List of Figures

Figure 1: Development of the US mortgage market 2001-2006

Figure 2: Example of a typical subprime MBS

Figure 3: New one family houses sold per month in the US

Figure 4: Funding of Northern Rock

Figure 5: Gross mortgage lending in the UK

Figure 6: Performance of FTSE 350 and FTSE 350 Banks in 2007

Figure 7: Distribution of events

Figure 8: Performance of FTSE 350, FTSE 350 Banks, and alternative portfolio

Figure 9: Federal funds target rate 2000-2008

Figure 10: S&P/Case-Shiller® Home Price Index 2000-2007

Figure 11: Development of the Dow Jones Industrial Index in 2007

Figure 12: Bank portfolio’s and alternative portfolio’s betas in 2007

List of Tables

Table 1: The top US subprime mortgage originators in 2006

Table 2: Relevant companies for the event study

Table 3: Categories of news about private financial institutions

Table 4: Categories of news about central banks

Table 5: Summary of the event study hypotheses

Table 6: CAARs of 3-day and 5-day event windows

Table 7: Timeline of the subprime-crisis in 2007

Table 8: CAARs after positive events for different estimation periods

Table 9: CAARs after negative events for different estimation periods

Table 10: CAARs after positive events for different event windows

Table 11: CAARs after negative events for different event windows

Table 12: Average abnormal returns after positive events by company

Table 13: Average abnormal returns after negative events by company

Table 14: Average abnormal returns after positive events by news category

Table 15: Average abnormal returns after negative events by news category

Table 16: Portfolio characteristics in 2006

Table 17 Additional constraints for the alternative portfolio

Table 18: Returns, betas, and weights of FTSE super sectors in 2006

Table 19: Characteristics of bank portfolio and alternative portfolio in 2007

List of Abbreviations

illustration not visible in this excerpt

1 Introduction

1.1 Motivation and Objective

The US subprime-crisis became a headline in the global media starting in February 2007 after the US housing market had already shown first signs of a slowdown in late 2006 (DiMartino & Duca, 2007, p. 1). Previously, the US housing market had enjoyed a favorable environment, especially from 2002 to 2005, which was characterized by low interest rates, rising house values, and increasing home financing possibilities through subprime mortgages (Krinsman, 2007, pp. 13-14). However, more and more events were published during the year by US mortgage brokers, international investment banks, and central banks around the world that presented a picture which caused today’s perception of the subprime-crisis. What’s more, the subprime-crisis is far from being over: an end to the crisis is not yet in sight.

One rather unique characteristic of this crisis is that its actual basis is the delinquencies and defaults of subprime single-family home mortgages in the US (Kiff & Mills, 2007, p. 3) which is commonly not regarded to be of great relevance for the international capital markets. However, taking into account the originate and distribute business model of US mortgage brokers in connection with the securitization of these mortgages into various types of securities that are traded on a global basis, it is not surprising to observe that banks and investment funds around the world were invested into these securities. Before the crisis started, only a few banks or funds considered the liquidity of these securities when investing significant amounts of money in them because they focused on maximizing their returns. But, when larger problems in the US subprime mortgage market became evident, liquidity became the major concern for investors and investor preferences significantly shifted to safer assets such as government bonds (Fender & Hördahl, 2007, pp. 9-11). This caused severe problems in the money market, which ultimately brought the crisis across the Atlantic to Europe. Moreover, funding problems emerged and caused the first bank run in Europe in decades when depositors in Britain started to queue outside Northern Rock branches for hours to withdraw their deposits in light of fears that the bank might have to file for bankruptcy (The Economist, 2007a, para. 28). In addition, another British bank had been in the spotlight earlier that year because HSBC was the first European bank to announce a billion dollar write-off linked to its exposure to subprime mortgages.

Taking into consideration the subprime-crisis-related events in Europe, the British banking market can be characterized as the only banking market in Europe where the subprime crisis caused banks to substantially write down subprime-related assets on the one hand but where severe funding problems even led to a bank run that had to be bailed out by the central bank and the government on the other hand. Consequently, the British banking market can be considered to be the European banking market with the highest impact of the subprime-crisis and is, therefore, worth analyzing in detail.

The objective of this thesis is to discuss the reasons for the emergence of the subprime-crisis and to empirically examine whether the subprime-crisis had an impact on the British Banking sector. The empirical analysis consists of two different approaches whereas an event study measures the short-term impact of certain news. The performance of the British banking sector in the full year 2007 is analyzed in a second approach that focuses on the long-term impact of the subprime-crisis. In addition, the paper provides an overview on the development of the subprime-crisis in 2007 based on a detailed description of the underlying fundamental market characteristics.

In order to empirically measure the impact of the subprime-crisis on British banks, an event study will be conducted. Event studies are a widely-used empirical methodology in economics and finance to examine the impact of certain events: they are considered to be the standard method to measure security price reactions (Binder, 1998, p. 111). An event study is an empirical study that measures if specific events have a significant impact on certain stock prices by calculating abnormal stock returns around predefined events. In this regard, an abnormal return is the difference between the actual return in the market and the expected return according to a return generating model (Peterson P. P., 1989, p. 36). A common assumption in this regard is that positive events lead to positive abnormal returns whereas negative events cause the abnormal returns to be negative. Consequently, important news relating to the subprime-crisis will be categorized as positive or negative and its impact on stock returns will be determined. The event study, as well as the timeline of the subprime-crisis, include events from January 1, 2007 to December 31, 2007. The analysis of the year-round performance of the British banking sector in 2007 is conducted in addition to the event study and follows a different methodology. In contrast to the analysis of the impact of individual events, this approach deals with the performance of British banks and compares this to the performance of an alternative non-bank portfolio. Key to this analysis is that both portfolios have the same risk and return characteristics at the beginning of 2007 that have been determined through a backtesting of the portfolios’ performance in 2006.

1.2 Course of the Investigation

In the second chapter, important fundamentals of the subprime-crisis will be examined. These fundamentals explain how an environment was able to develop in the last decades that lay the foundation for today’s crisis. In Chapter 2.1, an overview about the development and the structure of the US subprime mortgage market will be presented before specific characteristics of subprime mortgages will be outlined in 2.2. The unique business model of mortgage brokers is depicted subsequently. The last segments of Chapter 2 specify complex financial instruments that enabled the subprime-crisis to spread around the world and explain why the securitization process leads to high-risk securities.

Chapter 3 specifically describes the development of the subprime-crisis in 2007. After presenting an overview about the situation of the US housing market up to 2007 in 3.1, a timeline about last year’s subprime-crisis is outlined in 3.2, and the impact on international capital markets is discussed in 3.3. Chapter 3.4 focuses on the consequences for British banks and the actions of the British financial regulatory environment.

An empirical analysis of the subprime-crisis is conducted in Chapter 4. A general overview about event studies and their historic development is presented in 4.1. After deducing the typical framework of an event study in 4.2, the relevant British banks in line with its market index as well as relevant news for the event study are determined in Chapter 4.3. The actual event study that analyzes the impact of the subprime-crisis on British banks will be presented in Chapter 4.4. Additionally, a comparison of the performance of a bank portfolio with an alternative non-bank portfolio is given in 4.5

Finally, Chapter 5 contains a summary of the theoretical concepts and the empirical results and gives an outlook about a potential development of the subprime-crisis, capital markets, and specifically the British banking market. Ideas for further research are also presented.

2 Fundamentals of the Subprime-Crisis

2.1 The US Housing and Subprime Mortgage Market

Subprime mortgages are a common way to finance home ownership for people that did not have the possibility of getting a prime mortgage in the past (Dynan & Kohn, 2007, p. 26). These mortgages are loans for residential property which have a higher probability of delinquency and default than prime loans. Major differences regarding the characteristics of the loans are that upfront costs such as application and appraisal fees as well as continuing costs such as principal and interest payments are higher for subprime mortgages (Pennington-Cross, 2003, pp. 279-280). The differentiation between prime and subprime borrowers is usually made on the basis of the borrower’s credit rating which is usually determined by their credit score as well as loan ratios such as the debt service-to-income (DTI) ratio and the mortgage loan-to-value (LTV) ratio (Kiff & Mills, 2007, p. 3). The LTV ratio refers to the size of a mortgage lien as a percentage of the total house value (Brueggeman & Fisher, 2008, p. 145), whereas the DTI ratio measures the percentage of the borrower’s monthly gross income used to service the mortgage (Bhattacharya, Fabozzi, & Chang, 2001, p. 5). Fair, Isaac and Company established the widely used FICO credit score for mortgage borrowers that assigns a credit score to each borrower based on the probability that the borrower’s mortgage will default or that the borrower will become significantly delinquent (Frankel, 2006, pp. 69-70). These FICO scores range between 400 and 900 points (Taff, 2003, p. 121). Typical subprime mortgages are mortgages of borrowers with low FICO scores below 620, a DTI of more than 55 percent and a LTV of more than 85 percent (DiMartino & Duca, 2007, p. 4). Besides these rather objective criteria, it was discovered that borrowers who had subprime mortgages in the past are likely to only be able to receive subprime mortgages in the future even though their credit rating might have improved (Courchane, Surette, & Zorn, 2004, p. 381).

Mortgages with a higher credit quality than subprime mortgages but with a lower quality than prime loans are referred to as Alternative A loans (Alt-A) loans. These mortgages are often characterized as being obtained from borrowers who do not provide the mortgage lender with full documentation about their credit history and current income (Kiff & Mills, 2007, p. 3). In 2006, 81 percent of all Alt-A mortgages were sold without complete documentation, whereas only 50 percent of all subprime loans were originated without being fully documented (DiMartino & Duca, 2007, p. 4). Moreover, Alt-A mortgages are widely used by investors to finance property that will be rented to third parties (Ashcraft & Schuermann, 2007, p. 20).

The fourth mortgage category, which is in addition to prime, Alt-A, and subprime, is jumbo mortgages. In general, jumbo mortgages have the same characteristics as prime mortgages. The only difference between prime and jumbo mortgages is the size of the mortgage principal. Jumbo mortgages are defined as mortgages of prime quality with a principal above $417,000 (Ashcraft & Schuermann, 2007, p. 7). Due to the high level of similarity of prime and jumbo mortgages, this thesis will not examine jumbo mortgages in detail.

Kiff and Mills (2007, p. 3-4) present an overview about the historic development of the current US subprime mortgage market. The first factor that led to the emergence of the subprime mortgage sector was the imposing of interest rate caps in 1980. One major milestone was a change in the legal environment that introduced adjustable-rate mortgages (ARM) in 1982. The 1986 Tax Reform Act strengtend the subprime mortgage markets as the act abolished tax deductions of interest payments of all consumer loans except residential mortgages. As a consequence, residential mortgages were the only type of loans available to consumers of which the interest payments could be deducted from tax payments. Mortgage loans were now prefered to other forms of consumer loans for home improvements and other personal consumer spending because high-cost mortgages became effectively cheaper than consumer loans (Chomsisengphet & Pennington-Cross, 2006, p. 38). In addition to legal and regulatory changes that influenced the mortgage market, advances in the financial services industry also promoted subprime mortgages. Advanced computer models were developed, which led to a more cost-efficient mortgage origination process for banks. Moreover, standardized credit scores were introduced in the mortgage market after they were succesfully used in the auto loan market to provide an objective view on borrower’s credit risk (DiMartino & Duca, 2007, p. 2).

The securitization process, which will be described in detail in part 2.4, enabled banks to distribute credit risk off its own balance sheet and, consequently, led to the emergence of specialized mortgage brokers who offered mortgages without taking the total credit risk. The increasing demand for high-yielding residential mortgage-backed securitized products led to incentives for mortgage brokers to offer risky mortgages because investors were seeking high-yielding assets (Schumer & Maloney, 2007, p. 20). Moreover, the very low interest rate levels in the US which resulted from the burst of the internet buble in 2000 and the 9/11 terrorist attacks in 2001 made mortgages in general more attractitive as the Federal funds target rate, the US prime interest rate determined by the Federal Reserve System (Fed), decreased from 6.5% in 2001 to 1.0% in 2003 (Ashcraft & Schuermann, 2007, p. 7; Appendix A, Figure 9).

In line with the low interst rates during this time, housing prices across the US continously increased, enabling borrowers to refinance their mortgages and extract equity from from these loans (Chomsisengphet & Pennington-Cross, 2006, p. 41). These numeorous factors supported the establishment of the subprime mortgage market in the mid ‘90s and enabled more people, who had not been ab able to get traditional mortgage financing in the past, to own homes. Due to the rising house values and relatively low interest rates, the homeownership rate enjoyed tremendous growth in the US, with its highest value of 69 percent, in 2004 (Garriga, Gavin, & Schlagenhauf, 2006, p. 398). Real home prices increased by almost 85 percent from 1997 to 2006 and the S&P/Case-Shiller® Home Price Index shows an acceleration of the annual increases of home prices up to 2006 (Appendix A, Figure 10). The annual home price increase was 8.5 percent in 2001 and rose to more than 15 percent in 2005 (Schumer & Maloney, 2007, pp. 2-3). Along with a high homeownership rate, delinquencies of subprime mortgages were at very low levels during 2004-2005 (Krinsman, 2007, p. 13).

The market shares of different types of mortgage loans have significantly changed in the last decade. Most originated mortgages were so-called “prime conforming” loans until 2003 which were purchased by the government-sponsored enterprises (GSEs) Federal Home Loan Mortgage Corporation (Freddie Mac) and Federal National Mortgage Association (Fannie Mae) in accordance with its very strict conforming criteria and later on guaranteed, securitized, and resold to investors. The strict GSE criteria include narrow characteristics regarding the size of the mortgage and are only available to borrowers with a good credit score who can also provide full documentation of their income. GSE mortgages are capped at $417,000 for single-family houses and stringent criteria regarding DTI and LTV ratios and proof-of-income documentation apply (Kiff & Mills, 2007, p. 6). By 2006, more than 50 percent of all mortgage originations could not be purchased by GSEs due to a lack of accordance with its strict conforming criteria. In line with the declining market share of prime conforming loans, subprime loans were more and more originated. Subprime mortgages only made up about nine percent of the total mortgage market in 2001 and its market share stayed constant until 2003. But looking at recently originated mortgages presents a different picture. The share of subprime and Alt-A mortgages from all newly-originated mortgages increased from nine percent in 2001 to 20 percent in 2006 (DiMartino & Duca, 2007, p. 2). The following figure also shows that the percentage of subprime mortgages which have been securitized strongly increased from 50 percent in 2001 to 81 percent in 2006.

Figure 1: Development of the US mortgage market 2001-2006

illustration not visible in this excerpt

Adapted from Schumer and Maloney (2007, p. 10)

2.2 Characteristics of Subprime Mortgages

In recent years, a wide array of subprime and Alt-A mortgages were offered in the US. Almost all of these mortgages were hybrid loans that incorporated floating as well as fixed interest rates. Approximately two-thirds of these mortgages were so-called “2/28” mortgages that are structured as a fixed-rate mortgage (FRM) in the initial two year period and then revert to an ARM in the subsequent 28 years (Kiff & Mills, 2007, p. 8). The floating rate during the ARM period is usually reset every six months on the basis of a predefined benchmark such as the London Interbank Bid Offered Rate (LIBOR) or other benchmark interest rates plus a margin (Schumer & Maloney, 2007, p. 2). The margin is typically between 400 and 600 basis points (Krinsman, 2007, p. 15). The fixed rate is usually set below market interest rates to attract borrowers to the mortgage: it is therefore known as a teaser rate. Borrowers with these mortgages are subject to a substantial increase in interest payments when the mortgage converts from low fixed rates to higher floating rates. This characteristic can be diminished through the incorporation of interest rate caps but the rise of the interest rate is still substantial in most cases (Ashcraft & Schuermann, 2007, p. 22). Moreover, most subprime and Alt-A ARMs are structured as interest only (IO) and negative amortization (neg-am) mortgages in the first years of the mortgage. In IO mortgages, the borrower only pays the interest accruals in the first years and begins to pay down the principal after that period. In neg-am mortgages, the borrower only pays a part of the interest accruals in the first years. After this period ends, the borrower has to pay down the principal and additionally the interest payments including the remainder of the accrued interest of the neg-am period (Kiff & Mills, 2007, p. 8).

The level of interest rates for all mortgages depends heavily on the credit history of the borrower and to some extent the mortgage’s LTV. Borrowers with high FICO scores are regularly subject to the lowest interest rates. The interest rates for mortgages with similar FICO scores but different LTVs are relatively similar except for LTVs greater than 100 when significant premiums have to be paid by the borrower (Chomsisengphet & Pennington-Cross, 2006, pp. 48-49).

Another important characteristic of subprime mortgages are prepayment penalties. The share of subprime mortgages with prepayment penalties substantially increased in the last two decades and about 80 percent of all ARMs that originated between 2000 and 2002 include a prepayment penalty whereas this holds true for only 45 percent of all FRMs (Chomsisengphet & Pennington-Cross, 2006, p. 53). Prepayment penalties often last over a three or more year period and exacerbate the prepayment of mortgages as they can cost five percent or more of the mortgage balance. As a result, these borrowers cannot achieve lower home funding by refinancing their mortgage in times of decreasing interest rate levels or improvements in their credit ratings (Schumer & Maloney, 2007, pp. 21-22). The main reason for the introduction of prepayment penalties is derived from the fact that most subprime mortgages are securitized into mortgage backed securities (MBS). The value of these MBS is significantly affected by mortgage prepayments because prepayments cause changes in the duration of the MBS (Mason & Rosner, 2007, pp. 17-19).

2.3 Business Model of US Mortgage Brokers

The US mortgage market can be regarded as being separated by prime and non-prime mortgages. This separation especially applies to the institutions that are involved in these markets. The prime mortgage market is dominated by the GSEs because they define the prime conforming criteria which have to be met in order for prime loans to be purchased and guaranteed by the GSEs. In contrast, the non-prime market is regarded as being separate from the prime market, as subprime brokers usually focus on subprime loans only (Danis & Pennington-Cross, 2005, p. 6).

Most mortgage lenders in the US are specialized mortgage brokers who do not retain the mortgages they originate on their own balance sheet and, consequently, bear very little credit risk in case of mortgage defaults. Nevertheless, this makes them dependent on access to the market for securitized products because they do not have sufficient equity to fund mortgages themselves for a longer period (Alexander, Grimshaw, McQueen, & Slad, 2002, p. 671). Mortgage brokers have an especially strong market position in the subprime mortgage market where about 63 percent of all mortgages are originated through mortgage brokers (Kiff & Mills, 2007, p. 11). The subprime mortgage market is characterized as having a very high concentration of subprime mortgage originators.

Table 1: The top US subprime mortgage originators in 2006

illustration not visible in this excerpt

From Ashcraft and Schuermann (2007, p. 9)

In 2006, the top 10 subprime mortgage originators accounted for 60.3 percent of the total subprime mortgage market and the top 25 even made up 90.5 percent of the market. Looking at the top 10 mortgages originators, seven of them are specialized mortgage brokers and only three are normal commercial banks (Table 1).

As mortgage brokers do not take deposits, they are not subject to the strict regulations that apply to deposit-taking banks. Instead, they are only loosely regulated in accordance with the Home Owners’ Equity Protection Act and the Community Reinvestment Act (Schumer & Maloney, 2007, pp. 17-18).

The most important characteristic of the business model of mortgage brokers is the use of securitization to dispose the credit risk. In this business model, mortgage brokers deal with the application of the borrower and are responsible for the borrower’s credit check and the documentation and underwriting of the mortgages (Ashcraft, McAndrews, & Skeie, 2007, p. 10). They have a strong incentive to originate as many mortgages as possible without having the need to accurately check the quality of the borrower because they are paid on a fee basis that is linked to the number and value of the mortgages (Kiff & Mills, 2007, p. 11). This fee-based remuneration can even lead to predatory practices when brokers sell mortgages with unnecessary or harmful characteristics such as high prepayment penalties or higher interest rates than necessary with regards to the borrower’s default risk (Renuart, 2004, p. 480).

The commercial banks who were exclusively involved in offering mortgages before mortgage brokers were established do more in-depth checks of the credit risk of the borrowers because they either take the total credit risk on their own balance sheet or they transfer the risk to GSEs who have strict requirements regarding documentation and the credit score of the borrower. As a result of the increased market share of mortgage brokers, the mortgage’s default risk that used to be held by mortgage banks and normal commercial banks has been dispersed away from their balance sheets to the international capital markets.

2.4 Financial Instruments Underlying the Subprime-Crisis

One key aspect for understanding the current subprime-crisis is the transformation from subprime mortgages into new securities through the process of securitization. In general, securitization means that certain homogeneous assets that produce cash flows such as mortgages or bank loans are combined and put into a pool that serves as collateral for a new marketable security called asset-backed security (ABS). The debt service of the ABS is directly derived from the cash flows of the underlying assets (Graff, 2006, pp. 241-242). Consequently, the credit risk of the ABS only depends on the credit risk of the underlying assets and is, therefore, independent of the credit risk of the original owner of the assets (Breidenbach, 2005, pp. 35-41). The most common form of ABSs is MBSs, which can be backed by both residential mortgages and commercial mortgages (Fabozzi & Ramey, 2000, pp. 564-565).

The quality of MBSs can be increased through various measures to protect tranches with a high rating from losses of the underlying collateral. The most common measures are structural features of the MBS: subordination, overcollateralization, and excess spread (Kiff & Mills, 2007, p. 5). MBSs that are structured using subordination are divided into tranches which represent different classes of bonds ranging from senior tranches to junior tranches. These tranches can be distinguished by their credit risk and their return characteristics. Defaults of the underlying mortgages are applied to the junior tranches first and do not have an impact on the principal of the senior tranches until the principal of the junior tranches is completely exhausted. As a result, the junior tranches are characterized by high risk that is compensated by a higher yield in comparison to the senior tranches, which are very secure and, consequently, have a very high rating (Kiff & Mills, 2007, p. 5). The subordination feature that leads to different debt tranches is the reason why the senior tranches can be rated investment grade even though the underlying collateral consists of risky subprime loans (Buiter, 2007, p. 2). The general MBS consists of approximately 80% AAA-rated “Class A” tranches and 20% “Class M” mezzanine tranches that usually comprise 10% AA tranches, 5% A tranches and a 5% tranche with a rating of BBB+ and below. In addition, there is a small non-rated equity tranche that impairs the first losses of the underlying mortgages (Kiff & Mills, 2007, p. 5). This equity tranche is usually created through overcollateralization in a subprime MBS (Ashcraft & Schuermann, 2007, p. 34). An example of a typical subprime MBS structure is presented in Figure 2.

Figure 2: Example of a typical subprime MBS

illustration not visible in this excerpt

From Ashcraft & Schuermann (2007, p. 35)

Numerous investors, including pension funds, mutual funds, banks, and insurance companies invest in the “Class A” tranche because this tranche combines a high AAA rating of government bonds with a relatively high yield compared to a government bond with the same rating. These investors usually only invest in this super senior tranche and not in the mezzanine tranche. The major reasons for international investors to invest in securities which are based on US subprime mortgages are the diversification of their asset portfolio and the generation of higher returns these products offer in comparison to other investment products with an equivalent credit rating. Especially securitized products offer a wide range of possibilities to diversify the portfolio with regard to asset classes, regional aspects and risk-return-profiles (Altunbas, Gambacorta, & Marqués, 2007, p. 12). As a result, the “Class M” tranche is usually resecuritized into collateralized debt obligations (CDOs). The non-rated below mezzanine tranche is either kept by the MBS originator or sold to specialized hedge funds or proprietary trading desks of investment banks (Kiff & Mills, 2007, p. 5).

Another typically used possibility to enhance the credit risk of a MBS is overcollateralization, which can be achieved by structuring the MBS with a higher principal balance of the underlying mortgages in comparison to the par value of the MBS. Consequently, the first mortgage defaults are absorbed by the value of the overcollateralization before the MBS is impaired (Kiff & Mills, 2007, p. 5).

The last common structural credit enhancement feature is excess spread, which means that the weighted average coupon from the mortgage loans is higher than the coupon payments to MBS investors plus all other fees paid by the MBS. In other words, excess spread is a specific part of the interest payments of the mortgage borrowers that is specifically excluded from serving the collateral and that is used to increase the overcollateralization of the MBS (Kiff & Mills, 2007, p. 5). In 2006, average excess spread for subprime MBS was estimated to be 2.5 percent (Ashcraft & Schuermann, 2007, p. 36).

Besides these structural measures to enhance the quality of MBSs, external mechanisms including credit insurance can also be used. Specialized credit insurance companies which are also known as monoliners such as Municipal Bond Insurance Corporation (MBIA) and Ambac Assurance Corporation (AMBAC), offer protection for the senior tranche of MBSs and CDOs. They guarantee a timely payment of interest and the repayment of the principal. As a consequence, rating agencies rate the senior tranche AAA (Gangwani, 1998, p. 35).

As mentioned above, one important investor in the MBS market are CDOs which are almost the only investor in the mezzanine tranches but that also invest in high-grade “Class A” tranches. Residential MBSs are estimated to account for 40 percent of total CDO collateral (Mason & Rosner, 2007, p. 27) whereas this figure is estimated to have increased in the last years (Fitch Ratings, 2006, p. 1). Subprime MBS mezzanine tranches even made up about 70 percent of the total collateral in mezzanine structured finance CDOs in 2006 (Krinsman, 2007, p. 14). CDOs are investment vehicles that invest in a pool of assets such as bank loans or MBS. Similar to ABSs, CDOs also fund their investments by issuing different classes of securities and its returns are linked to the returns of the underlying assets (Schorin & Weinreich, 2000, p. 241). But unlike ABSs, most CDOs are actively-managed by specialized investment managers who invest its funds according to predefined covenants that clearly state the average ratings of the invested securities and the share of securities with a certain credit rating (Ashcraft & Schuermann, 2007, p. 15). As a result, these CDOs can be forced to sell certain MBS tranches when these tranches are downgraded by credit agencies.

Although the mezzanine MBS tranches only make up 10 percent of the total MBS value, Mason and Rosner (2007, pp. 27-28) argue that CDO investments in “Class M” tranches are critical for the entire MBS market because all “Class A” tranches cannot be sold without the mezzanine tranches being sold as well. Therefore, they conclude that CDO demand for MBSs has a leveraged effect on the possibilities for mortgage securitization and, consequently, this leveraged effect transfers over to the availability of residential mortgages. As a result, a decrease in investor demand for mezzanine subprime CDOs can have a huge impact on the subprime mortgage market itself. In addition, this effect is increased by similar characteristics of the CDO market itself, where the mezzanine tranches are mostly sold to so-called CDO2s. This refers to CDOs who themselves invest in other CDOs (Watterson, 2005, p. 6).

Another financial instrument that became evident during the subprime crisis was asset-backed commercial papers (ABCPs) which are often used in connection with off-balance sheet conduits. ABCPs have a maturity of up to nine months and used to be issued to securitize commercial trade receivables (Herrmann, 2005, p. 46). In the last years, banks started to set up arbitrage ABCP conduits that have a similar investment strategy as CDOs as they also invest in assets with a long maturity such as CDOs or MBSs. But unlike CDOs which are refinanced by long-term debt, these conduits are refinanced by short-term ABCP. Consequently, these ABCPs need to be continuously rolled over to ensure the funding of the assets. In order to continuously guarantee the funding of the conduits, the issuing bank has to provide liquidity backstops to the conduits (Fender & Hördahl, 2007, p. 7). These liquidity backstops were used by the conduits to repay its commercial paper holders at the end of the paper’s maturity when the conduit is not able to issue new ABCP to roll over the funding (Barr, 2007, para. 26). Moreover, ABCP conduits are used by financial institutions, especially mortgage brokers, to finance the accumulation of originated mortgages until they are securitized to MBS (Stone & Zissu, 2005, p. 154). The global ABCP market was estimated to have a size of $1.5 trillion at end-March 2007, of which $300 billion were regarded to be backed by mortgage-related assets (Fender & Hördahl, 2007, p. 7). Similar to an ABCP conduit, a structured investment vehicle (SIV) uses medium-term notes in addition to ABCPs as a funding tool (Fender & Hördahl, 2007, p. 8).

2.5 Consequences of the Fragmented Securitization Process

All participants in the value chain of subprime MBSs have limited incentives to be risk-avers, largely due to the fee-based remuneration that all of them receive. The historically strong relationship between mortgage lenders and mortgage borrowers is now fragmented between numerous parties along the mortgage lending value chain of which each of them have individual interest to maximize their profit (Krinsman, 2007, p. 14). Both subprime mortgage brokers as well as the investment banks that structure MBSs and CDOs receive major parts of their remuneration based on the volume of their products which leads to a focus on generating volume with a large number of deals instead of generating fewer high quality securities. This volume-based remuneration in combination with the dispersion of the credit risk through the securitization to a large number of investors reduces incentives to monitor quality in each part of the process (Kiff & Mills, 2007, p. 7).

Rating agencies play a key role in the market of securitized products as they assign credit ratings to MBSs and CDOs based on a cash flow simulation that takes into account the probability of default of the underlying mortgages and the measures of credit enhancement of the structured product. As described above, MBSs consist of thousands of individual mortgages and CDOs invest in numerous MBSs and, consequently, consist of even more individual mortgages, sometimes hundreds of thousands of mortgages. An MBS or CDO investor is therefore not able to establish his opinion about the quality of these products by himself in light of the complexity of the product and the underlying intransparency (Ashcraft & Schuermann, 2007, pp. 41-45). Instead, he has to use credit ratings as basis for his investment decisions. Consequently, these ratings are imperative for originating these products because they provide investors with an independent and objective view about the credit risk. Nevertheless, credit agencies are also remunerated on a fee basis and, therefore, are keen to continuously advise and rate new securitized products (Kiff & Mills, 2007, pp. 11-12).

3 The Development of the Subprime-Crisis

3.1 Situation of the US Housing Market up to 2007

The strong momentum of the US housing market observed in the first half of this decade significantly changed in the last year. Beginning in late 2005, new house developments and house sales decreased. Building permits for single family houses declined by 52 percent from September 2005 to October 2007 (DiMartino & Duca, 2007, p. 1). Moreover, sales of new single family houses consciously declined from the end of 2005 to the end of 2007 (Figure 3).

[...]