The Rampant O & D Model
In highlighting the unique features of the current fi nancial crisis, it might be more important to identify and study the various institutional factors that enabled the fi nancial bubble to grow for a long time, than to study the fac- tors that triggered its bursting. More concretely, these are the factors that changed a problem in the US local subprime loan market in a blink of the eye into a problem for globally active major banks and for the global fi nan- cial system itself. As explained in the previous chapter, securitization based on the O & D model played a very important role in this process. In the fol- lowing, I will discuss why it aggravated the problem.
The traditional business model of banks is to originate loans and keep them on their balance sheet up to maturity. During this process, banks are supposed to manage risks associated with these loans. This model is the so - called “ originate and hold (O & H) model. ” By contrast, the O & D model is fi rst to originate loans but then to pool them, so that they can be sliced and diced according to the risk preference of markets, and then to distribute them to market participants such as institutional investors (see fi gure 2.1 ).
With this model, banks could profi t from commissions related to loan origi- nations while transferring the loans ’ risk to other parties.
The greatest merit of this model is to enable the risk existing in the fi nancial system to be transformed through securitization into transferable small pieces with various risk characteristics in accordance with various needs and capacity for risk taking of economic entities. Previously, the bank had held all the risks arising from loans. Even if other parties took on the risk of loans originated by the banks, the banks simply sold the loans to them without any transformation of the risk characteristics.
This new business model could satisfy the various risk appetites of non - banks or third parties, including institutional investors. For example, certain investors might hope to take on smaller risks in a certain industry than those involved with bank loans. For this type of investor, securitization could pool the loans extended to the industry, keeping the characteristics of risk of this industry (in terms of correlation with other industries) but
Real economy (household, corporate)
Debt overhang of household sector Unwinding of Yen carry trade and its impacts on forex/share prices
Banks Securitization
Rating agencies
Mortgage companies Fund raising Liquidity crisis Liquidity enhancement
Credit enhancement
Separation of originators and holders, Increasing complexity of risk structure, trade shifting to non-banks Increase in information asymmetry
Rapid contagion of market anxiety through information asymmetry
• Efficient financial intermediation
• Wider risk taking options for household and corporate
• Increase in risk trading activities, which leads to optimal risk allocation
• Increase in robustness of household and corporate against risks
• Enhancement of financial institutions’ profitability
Regulators/Central bank
Market participants Direction of flow of funds
on cies
Figure 2.1 The O & D model
tranched them into different risk categories (e.g. in terms of probability of default), thereby creating a modern monster with, for example, characteris- tics of the credit risk of the real estate industry, but with a lower probability of default than would be typical of loans to this industry.
Furthermore, loan pooling could reduce the size of risk compared with a simple aggregation of the risk of each loan. This is because the probability of simultaneous default is generally supposed to be very minimal. In particu- lar, if we can capture the default correlation, that is, the probability that the default of a certain loan occurs simultaneously with the default of another, we can use this pooling effect in an effi cient way to reduce risk.
Consequently, the O & D model dramatically increased the system ’ s ability to take on risk. Owing to securitization, the risk previously taken principally by banks was transformed into small parts with various risk characteristics, so that many investors could take only their preferred type and amounts, leading to a system that could take a bigger risk as a whole than before.
There are few academic studies that analyzed the positive impacts of this business model on the macroeconomy but, for example, Hamilton, Jenkinson, and Penalver (2007) cited the following points:
■ Benefi ts to the corporate sector:
– an increase in the number of corporations that could directly access fi nancial markets (see fi gure 2.2 );
Overview of the Financial Crisis 21
– easier access to the long - term credit market;
– easier control of risks through various derivatives;
■ Benefi ts to the household sector:
– an increase in alternative mortgage loans and the increase in the maxi- mum loan amounts that can be borrowed thanks to the advancement of banks ’ risk management techniques;
– expansion of access to various credit markets (see fi gure 2.3 );
■ Benefi ts to the macroeconomy:
– higher growth rate of the fi nancial sector in major countries and a subsequent increase in the share of the fi nancial sector in the econ- omy. (As an example, this ratio increased in the UK from 5 percent in 1985 to 7 percent in 2006.)
So the O & D model was generally accepted as a symbol of fi nancial and macroeconomic development that could never be reversed, particularly in countries such as the US and UK, where this model was rampant. Also, the basics of this idea itself have been retained in many countries, even after some large fi nancial institutions collapsed and the authorities of many US and European countries were forced to inject public money into the banks.
0 0
2006 No
AAA AA A BBB BB B C–CCC
No 5000 4000 3000 2000 1000 5000
4000 3000 2000 1000
2000 1994
1988 1982
1976 1970
Figure 2.2 Number of corporates assigned ratings
Source: Hamilton, Jenkinson, and Penalver (2007), based on material from Moody ’ s
0 10 20 30 40 50 60
0 10 20 30 40 50 60
2006
% %
2002 1998
1994 1990
1986
Figure 2.3 The share of individuals who hold credit cards in the UK in credit outstanding
Source: Hamilton, Jenkinson, and Penalver (2007), based on material from the Association for Payment Clearing Services, Bank of England, National Statistics
Failure of the O & D Model: Expansion of Information Asymmetry
The next question naturally is why such a rosy model eventually caused the current fi nancial crisis. The answer is simple; it is mainly because major economic entities such as banks, investors, rating agencies, regulators, and the government failed to share information on the risk.
As stated, the O & D model expanded the number of entities that could share the risk, which was previously taken only by banks, increasing all of society ’ s capacity for risk taking. An important element in this model ’ s per- formance, however, is whether the risk information is properly transferred when the risk is transferred, or whether this model is properly equipped with a system that enables this risk information to be properly transferred.
In other words, this is a question of the degree of tradability of risk or of socialization of risk. Tradability of risk here indicates a situation in which sharing risk information among market participants widely facilitates stable transactions in the market, and socialization of risk indicates the situation
Overview of the Financial Crisis 23 in which a certain well - established consensus exists in the society on how to share risk information among concerned parties.
Needless to say, fi nance is the business of dealing with information asym- metry among various economic entities. Information asymmetry is defi ned in economics as “ a situation in which the parties engaged in a certain trans- action have different levels of information about the transaction. ” The party with inferior information might act with an increase in uncertainty because of the lack of information (if it recognizes this situation), which might lead to the failure of the transaction or a price far away from the one that would be obtained without any information asymmetry. This is one typical cause of “ market failure. ”
A typical example of asymmetry in the banking business is that which arises between banks and obligors in information on the obligors ’ credit- worthiness. Particularly, obligors with bad creditworthiness have an incen- tive to hide this inconvenient information from banks when they borrow money from them. For this reason, banks naturally need to devise measures to reduce this information gap, or to set up lending conditions that discour- age obligors from abusing the asymmetry.
Similarly, information asymmetry occurs between originators and investors in the process of securitization. Originators have an incentive to transfer bad - quality products to investors, capitalizing on their advantage in information. Ordinarily, the rating agencies ’ examination of the securities and assignment of ratings to them, or disclosure of information related to securitizations, or various conventional practices and regulation and super- vision by the authorities are expected to contain the possibility for abuse.
Tradability of risk or socialization of risk is only ensured by establishing a mechanism that supports the proper functioning of all these means of containment.
In the current fi nancial crisis, this containment failed to function well for various reasons, thus leaving the crisis to go out of control. Why could they not work to stop the fi re from spreading? An answer can be found in the speed of the expansion of information asymmetry in the past few years, which was too rapid for these means to contain.
A typical example of this expansion of information asymmetry is the introduction of highly complex securitization products. For example, rese- curitized products with underlying assets that are themselves securitiza- tion products, such as ABS of CDO or CDO squared, became popular and consequently have been sold widely in the market in recent years (see fi gure 2.4). The risk associated with these products, however, was estimated on the assumption that the risk associated with the underlying assets was correctly understood. In this process, there seems to have been no validation of the
Bank, mortgage company
Mortgage loan
ABCP
RMBS ABCP AAA
AA BBB No rating ABCP
AAA/
AA of RMBS,
etc.
ABCP
SIV
AAA/
AA of RMBS,
etc.
ABCP/
Senior debt
CDO
Senior/
Mezzanine of RMBS,
etc.
Others Subordinate debt
AAA
AA BBB No rating Mortgage
loan Mortgage
loan Senior parts of securitization products
Figure 2.4 The resecuritization process
Overview of the Financial Crisis 25 risk associated with the underlying securitization products and no consider- ation of model risk.
Generally speaking, any estimated risk tends to be quite sensitive to a variation in assumptions, so some model risk is hard to avoid. This type of risk can sometimes be quite substantial, so a capital buffer is set aside for it.
The risk of securitizations of securitized products is likely to be underesti- mated if no model risk is considered. In other words, ignored model risk is multiplied with each increase in layer of securitization. In the current fi nan- cial crisis, these securitizations were the ones that suffered most.
Another aspect of the expansion of information asymmetry in the recent securitization market is the increasing share of non - banks in the US mortgage market. Also in Japan, we have seen some non - banks recently entering this market, but in the US they started to enter it in the early 1980s. Particularly in the US subprime loan market, non - banks had already become a major player.
For example, Norinchukin Research Institute (2002) indicated that the share of deposit - taking fi nancial institutions in the mortgage loan origination mar- ket was about 70 percent in the early 1980s, but fell to about 50 percent at the end of 1990s. Meanwhile, in the same period, mortgage companies ’ share increased from about 20 percent to about 50 percent. Furthermore, deposit - taking fi nancial institutions ’ share of mortgage loans held after origination dropped to about 30 percent in recent years (see fi gure 2.5 ).
0 20 40 60 80 100
1978 1982 1986 1990 1994 1998 2002 2006
0 20 40 60 80 100 Depository institutions GSEs (incl. securitizations) Private label securitizations REITS and households (percent of mortgages outstanding)
Figure 2.5 Declining mortgage holdings of deposit - taking fi nancial institutions in the US
Source: FRB
An emerging concern in this environment is the expansion of informa- tion asymmetry between the authorities and non - banks on the non - banks ’ origination practices. Unlike banks, non - banks were not usually subject to direct supervision by the authorities, so it was hard for the authorities to monitor them, even if there had been a signifi cant loosening of origination standards. Moreover, the authorities might have faced diffi culties in conduct- ing macroprudential policy without having the correct micro - information.
To be sure, US regulators had already published their warning about excess loosening of loan conditions in the subprime loan market in March 2007 (OCC 2007). In hindsight, however, this did not have enough impact on fi xing the problem.
The Mythifi cation of Risk Management
The second cause behind the development in the O & D model that led to a serious crisis lay in the overconfi dence in various risk management tech- niques to overcome the information asymmetry that arises in securitization transactions. This overconfi dence led to many oversights of serious risk management weaknesses.
In the previous section we observed a recent rapid increase in information asymmetry in the securitization market. We then need to consider how the concerned parties deal with the risk associated with this information asym- metry. For example, fi nancial institutions ’ risk management techniques were supposed to be suffi ciently advanced to analyze the complex risk associated with securitization transactions precisely. However, when doing advanced risk management like this, we often forget the basics rather easily.
For example, model risk was often ignored when estimating the risk of securitizations of securitization products. Besides, although it is not only an issue of securitization, many fi nancial institutions simply followed the out- come of VaR without asking further questions. VaR is supposed to indicate the probability of occurrence of huge losses using a statistical model based on the volatility of risk factors observed during a certain period.
This issue related to VaR will be discussed more in detail in chapter 4 . The one thing to be noted here is that VaR itself is not problematic, but this way of using it is. Even though there are many assumptions in a VaR calcu- lation, many experts easily forgot them and simply used its outcome as an unequivocal answer. This behavior caused the problem.
Also, many fi nancial institutions and investors buying securitized prod- ucts substantially outsourced their own risk analysis of these products to third parties, that is, rating agencies. In other words, they simply followed the ratings assigned by rating agencies in assessing the risks of these prod- ucts and decided their investments on that basis.
Overview of the Financial Crisis 27 This type of investment style itself cannot necessarily be blamed. Rather the issue is whether investors managed the “ risk of outsourcing ” well, or whether they were well prepared for the situation in which the outsourcee made the wrong decisions. Unfortunately, many investors, including large fi nancial institutions, tended simply to follow the judgments of rating agen- cies and failed to review their assessments for themselves. This can be seen as a result of them putting too much stock in the rating agencies’ capability to assess risks.