VI. Effects of securitisation on the loan portfolio composition (loan book), credit risk exposure, asset funding of banks and banking regulation.

A. Regulatory Change and Its Effects

Loan securitisation harnesses the adversity of both the current one-size-fits-all regulatory straightjacket and the competition in lending markets, which renders the cost-effective origination of loan for the bank portfolio (especially of investment-grade credits) increasingly difficult. This predicament has prompted banks to consider balance sheet restructuring for purposes of mitigating regulator), capital as well as improving overall economic efficiency (Punjabi and Tierney, 1999).

The main channel through which banks arbitraged the regulatory provisions of the 1988 Basle Capital Accord was by securitising their better quality assets and retaining their riskier assets on their own books. Barring future modifications by the Basle Committee the equitable treatment of risk categories under the Capital Accord of 1988 (i.e. a constant capital risk weighting, which does not distinguish between different qualities of loans) still represents a perennial source of regulatory and institutional arbitrage. Consequently, the market for securitised assets grew dramatically from the early 1990s onwards and attracted a large following with all major investment banks for purposes of obtaining capital relief; gaining liquidity or exploiting regulator), capital arbitrage opportunities in the securitisation of loans. Since it is less efficient for banks to retain highly rated loans due to their tight spreads relative to the regulatory capital requirement (unlike high-risk loans with an interest sufficiently high to sustain a flat capital charge), the indiscriminate risk-weighting of loans has led a growing number of national and regional banks to concentrate on the securitisation of investment grade credits, whose inefficient relationship between associated regulatory capital requirements and interest yield constitutes an arbitrage opportunity. Only banks with a developed trading portfolio capability are in the position to remove credit risk of non-investment grade loans from their loan books as a result of this disparity between the regulatory regime and the economics of financial intermediation governing the benefits from loan business.

With the new proposal of the 1988 Basle Accord suggesting the implementation of discriminatory risk-weightings across rating categories, the prospective change of the current regulatory regime will censure institutional arbitrage on regulatory capital requirements, which has hitherto motivated asset-backed securitisation. The new proposal of the Basle Committee incorporates advances in credit risk measurement, as it allows minimum capital requirements for credit risk to be determined by an internal ratings-based approach (IRB). Consequently, different loan grades will attract different commensurate risk weights in the future, e.g. low credit risk of investment grade loans is transposed into a lower level of regulatory capital. If the previous broad-brushed regulatory treatment of loans rules out arbitrage opportunities of low-risk assets under the current risk-based regulatory framework, banks are very likely to dispense with investment grade loans at large in securitisation transactions. (45)

Conversely, as a higher capital charge levied on risky assets will carry larger risk-based capital haircuts, the incentive to securitise non-investment grade loans will rise. The relationship between the risk level of non-investment grade loans and the associated economic capital cost will determine the extent to which banks and other financial institutions are prepared to substitute high-risk assets (i.e. non-investment grade loans with presumably high capital haircuts) for investment grade-related credit exposures on their loan books--a reversal of the present drainage of low-risk loans off the balance sheet. Hence, loan securitisation, originally devised as remedy to inflexible regulatory capital charges, will be instrumental in the efficient management of economic capital for purposes adequate asset allocation. Therefore, the erosion of regulatory arbitrage by means of replacing the present regime of one-size-fits-all risk-based capital requirements is intimately related to improvements in credit risk management of banks and financial institutions.

Although the latest Basle proposal aims to moderate future regulatory incentives of banks to dispense with low-yielding assets through securitisation on an excessive scale, the market is now too large and important just to disappear. The unabated popularity of asset-backed securities raises some complex questions about how such securitisation should be treated for risk control purposes. The envisaged scrutiny of internal credit risk assessment presented in the new Basle Accord does not only probe a comprehensive examination of the bank-based computation of capital requirements of loan books as to the explicit treatment internal rating mechanisms. It also warrants contemplating the development of financial intermediation with respect to loan securitisation. This is a difficult question, especially since securitisation can be structured in a wide variety of ways, eventuating disparate risk profiles for both the originating bank and capital market investors. Unless rules on risk management, transparency and investor protection prove adequate, such form of structured finance could possibly pose a significant threat to the stability of financial markets.

While the benefits from regulatory arbitrage on investment grade loans fade in view of the new proposal to a new Basle Accord, the new reality of a more responsive regulatory setting does not invalidate but rather strengthen the argument of risk-adjusted efficiency gains (of economic capital) in the process of loan securitisation. Securitisation maintains its economic edge, as it enables banks and non-bank financial institutions to reap the rewards from advanced approaches in controlling credit risk and reduce inessential non-interest rate expenses.

B. Changes in the Configuration of Securitisation (46)

1. Standardisation 

The growing standardisation of loan terms and credit scoring processes does not only lead to operational efficiency and transparency of credit risk management routines but also fosters mitigation of inherent uncertainty in both the estimation of the cumulative distribution function of default probabilities and loss severity associated with various loan pools. Simulation models to estimate the performance of the reference portfolios of synthetic and conventional loan secutitisation as well as improved analytical systems for the credit risk assessment of portfolios, such as KMV's Portfolio Manager, address much desired properties of credit risk management. Higher precision in the estimation of credit risk (i.e. a declining marginal increase of total variance of estimates as expected losses rise) is tantamount to reduced credit risk exposure to unexpected loss.

Given the inherent complexity and diversity of structured transactions, Burghardt (2001) states that a case-by-case basis evaluation of structured products with a derivative element (46) (such as synthetic CLOs) or pure derivative transactions is inevitably warranted from both a risk and regulatory perspective. Therefore, greater transparency of credit risk through standardisation bodes well with the conservative procedures of rating agencies in the determination of default probabilities and the pricing of synthetic asset-backed securities. So far, especially in cases of new types of reference portfolio assets (most prominent in synthetic CLO structures), relatively low structured ratings for mezzanine tranches (intermediate credit tranches) have resulted in spreads well above those found for comparably rated corporate bonds with arguably lower uncertainty about asset quality. The proposed regulatory framework, however, instils greater efforts in closing the information gap between issuers of CLOs and rating agencies due to a greater degree of transparency and standardisation of credit risk assessment by means of second-generation models of credit risk analytics.

Although credit rating agencies as the prime source of credit risk analysis for CLO transactions will not be rendered redundant, the increase in bank-based credit risk assessment is most likely to improve the efficiency of CLOs. This, in turn, allows investors to draw comfort from an increased understanding of the credit risk inherent in CLO transactions (as informed buyers), whose diversity and complexity tends to cause problems in analysing the risk-return relationship for loss of appropriate analytical approaches (Burghardt, 2001), which could result in incorrect classification and underestimation of risk exposure.

As opposed to the notion of portfolio diversification, which redistributes risk by pooling numerous underlying asset risk return profiles, synthetic securitisation is predicated on the exclusive transfer of credit risk without renouncing loan servicing. This form of risk redistribution is particularly sensitive to credit risk sophistication of informed buyers. Investors would no longer deal with structured products in an undifferentiated way, unless breaking down structured products into individual risk elements imposes disproportionate resource cost, such as time and specific asset knowledge. If increased confidence stimulates informed demand for structured products as the information premium decreases, spreads decline and synthetic CLO structures become more attractive as modes of loan securitisation.

2. The structural make-up of loan securitisation

In the light of the proposed revision of the Basle Accord, the increased focus of securitisation on the efficient use of economic capital in lending business is strongly intertwined with the type of securitisation contingent on in-house credit risk management capabilities. In order to represent credit risk more truthfully for purposes of mitigating the internal ratings based capital charge, private placements with other financial institutions would no longer warrant major involvement of rating agencies. Hence, banks might be in the position to do without rating agencies in conducting securitisation transactions to fine-tune the composition of the loan portfolio (Punjabi and Tierney 1999).

Concurrent to the adoption of the internal-ratings-based approach, rising sophistication in credit risk management also implies an altered logic of the structural make-up of loan securitisation. With the mechanism of removing loans from the balance sheet through true sale being doomed to obsolescence (in absence of regulatory arbitrage), the creation of perfected security interest of a synthetic claim on the underlying reference portfolio becomes the method of choice. Former disincentives of synthetic loan securitisation--inadequate credit risk assessment and information disclosure--have grown devoid of much of their economic relevance as regulatory consideration of internal credit risk assessments rewards the close alignment of economic and regulatory capital. As the legal treatment of the servicer of a loan pool no longer constitutes regulatory benefits associated with true sale, the administration of securitisation appears to be best served by the arrival of "synthetication" of asset claims, which has stolen a march from traditional securitisation. Thus, provided that the migration towards a responsive regulatory system perpetuates the sophistication of credit risk management and rectifies arbitrage behaviour to hitherto defunct regulatory provisions, the emphasis on economic capital as the prime incentive of synthetic securitisation is essentially a child of its own making. The implications of regulatory change and advanced credit risk methodologies confine the optimal structure of loan securitisation to the transfer of credit risk only.

Although securitisation facilitates the cost-effective utilisation of economic capital, its economic benefit, however, varies across banks, depending on the varying degree of individual composition of loan portfolios and the economic objectives banks intend to achieve through securitisation. Conventional CLOs cater to issuers, who seek to allocate credit risk more equitably to investors, reduce the cost of capital of loan origination, and curtail balance sheet growth of the loan book. In some instances the issuer might simply not be adept in completing credit derivative transactions in compliance with commonly accepted regulatory principles and standards of credit risk control. In contrast, synthetic CLOs are widely revered for their capacity of efficient credit risk transfer instead of a clean break of credit-linkage through a true sale of assets. "Synthetication" represents a seachange in bank-based financial intermediation due to increased efficiency in economic capital, which results from converting individual, illiquid financial assets into tradable market instruments by means of combining debt securities and credit derivatives as financing conduits. Although the novel features of "synthetication" rebound in slightly wider spreads and marginally higher risk-based capital haircuts (Punjabi and Tierney, 1999), the relative ease of completing credit default swaps and a rapidly tapering learning curve of capital markets about "synthetication" permit structural flexibility of synthetic CLOs, whilst the servicing function of the reference pool of loans remains untouched.

3. Credit rationing and operational efficiency

Two consequences emanate from the prevalence of synthetic CLO security design. For one, synthetic CLOs garner issuers with a wide range of eligible assets for portfolio selection, beyond the conventional restriction to illiquid and fairly standardised loans in traditional securitisation. As much as the expansion of securitisable asset properties signals the perennial dynamics in structured finance, it coincides with an activist sprawl of standardisation in loan origination (e.g. credit scoring systems) by financial institutions. At the same time, the extended scope for asset selection in "synthetication" curbs Sears about credit rationing of non-standardised loan contracts, while mitigating the impending cost premium of non-standardised loans. Barring new banking book regulations concerning credit derivatives, also the consideration of non-loan risk, i.e. risk exposure unrelated to the reference portfolio but associated with the collateralisation of the issued debt securities (e.g. counterparty default on a credit default swap securing the super-senior investor tranche), augments the scope of application of "synthetication", with banks seeking to free up economic capital locked up in asset management provisions.

In general, the effect of securitisation per se on capital provision is straightforward. Loan securitisation espouses the basic concept of a more efficient use of economic capital (see Exhibit 22) and stretches asset funding beyond what would have been attainable by means of self-funding in traditional on-balance sheet lending due to the expansion of funding sources (besides ordinary account deposits).

[ILLUSTRATION OMITTED]

In keeping with the concept of risk diversification in modern portfolio theory, the ability to incorporate and sustain larger loan pools allows issuers of collateralised loan obligations (CLOs) to lower their overall credit risk (if we drop the assumption that issuers want to reduce balance sheet growth). The particular security design of CLOs allows issuers of CLOs to slice and dice the reference portfolio of loans according to estimated default by means of subordinating debt securities (various tranches with different seniority).

If asset proceeds and credit defaults are prioritised according to seniority (i.e. subordination through loss cascading), securitisation achieves a close match of the term structure of each tranche with the default tolerance of each risk type of investors in debt securities. The reconciliation supply and demand of risky asset claims commits less economic capital to the loan origination process and mitigates the potential for reduced non-interest expenses, as no unexpected credit risk should go unchecked in the optimal case of optimal market equilibrium under perfect risk classification. Consequently, the cost of administering securitisation transactions should be more than offset by economic gains derived from removing credit risk off the balance sheet. The level of trade-off (hurdle rate of securitisation) stands to be measured by the opportunity cost of interest proceeds commensurate to the asset quality of the securitised loans under information asymmetry.

4. Market mechanism and risk allocation

Optimal allocative efficiency (through regulatory arbitrage) does not necessarily equate to lower systemic credit risk. In fact, regulatory recognition of closer approximations of credit risk leaves little room for other risks impacting on banking business to be accommodated in regulatory capital requirements, such as operational risk (Goodhart, 2001). Hence, mechanisms of regulatory arbitrage for purposes of fine-tuning a previously broad-brushed determination of capital charges represent a most welcome market reaction if the means of achieving capital relief lead to an efficient allocation of capital, with risk being adequately diversified.

In the recent past, so-called monoline insurance (an insurance company set up with the sole purpose of guaranteeing selected tranches of asset-backed securities) has been a popular method of credit risk transfer for issuers in loan securitisation. Insurance companies guarantee to make good on credit loss of a pool of loans underlying a loan securitisation and, thus, free issuers from retaining minimum capital requirements for these loans. While this mechanism allows banks, for instance, to arbitrage present regulatory provisions and originate more loans, central bankers would not necessarily object to such techniques, which move risk away from banks, for which they may have to provide liquidity, as opposed to insurance companies, for which they will not (The Economist, 2002a). The original intention of risk diversification and allocative efficiency in regulatory arbitrage through credit risk transfer, however, has alerted financial watchdogs, who worry that an alignment of economic and regulatory capital through misguided credit risk transfer might lead to a build-up of risk elsewhere, or may not have been perfectly passed on to counterparties in derivative transactions (The Economist, 2002a).

Regulatory capital relief cannot sidetrack from the prospect of a dangerous reshuffling of individual credit risk exposures between financial service firms. Given that regulatory arbitrage of credit risk through third-party insurance is only acceptable if regulations imposed by national supervisors reflect different economics, i.e. any transfer of credit risk from banks to insurance firms requires the different objective and investment horizon of the counterparty to be a better match for the type of risk transferred. Rule (2001), however, states that little knowledge of insurers about the characteristics of loans and other debt transferred, or hedged, by banks invalidates the claim of increased efficiency and diversification through risk transfer. As banks and insurers treat credit risk differently, the absence of comprehensive information about individual credit risk in the loan pool might not reduce economic and regulatory cost. Hence, the rationale of credit risk transfer both from a firm perspective as well as industry perspective would have been rendered meaningless. Consequently, judging the feasibility of risk transfer in the context of structured finance boils down to how well it produces more efficient levels of regulatory capital issuers are prepared to provide. Any transfer of credit risk should be based on enhanced credit risk management, which must not reflect institutional arbitrage but a continued effort to allocated credit risk as efficiently and equitably within and across financial institutions as possible.

As much as the credit enhancement of CLO transactions, generally a structural sine qua non, predicts the first loss provision for estimated credit events reasonably well, regulators and banks are faced with the question of how the collateralisation of senior tranches through monoline insurance should be treated in terms of minimum capital requirements to reach similar regulatory outcomes as in the case of credit enhancement. Although issuers of securitisation transactions correctly estimated future losses and provide commensurate capital cover, the edifice of asset-backed securitisation in general and CLOs in particular heavily depends on the credit rating issued by rating agencies upon assessment of extreme cases of credit events; and so does the valuation of monoline insurance, which has been created to sustain high levels of structured ratings of securitisation transactions. Considering the doubtful default protection of such insurance in severe portfolio distress, the possibility of misallocated credit risk through regulatory arbitrage subjects more and more off-balance sheet financial activity in structured finance with third-party insurance cover to comprehensive credit risk assessment by rating agencies.

The reliance on such external ratings for purposes of averting misguided allocation of credit risk to insurance companies does not only boost the governance of CLO transactions by rating agencies and their interpretation of credit risk. It might also lead issuers, regulators and investors to fall victim to collective myopia that blinds them to the actual risks of what is being packaged into the reference portfolio of CLOs and asset-backed securitisation transactions (The Economist, 2002a). Therefore, the pervasiveness ofstandardised rating approaches applied in structured finance could reverse efforts of efficient risk diversification unless incentives of regulating securitisation coincide with the economic reality of the issuer's capabilities to manage credit risk--be it a bank or an insurance company.

5. Implications for bank lending

The attractiveness of securitisation, however, is not devoid of implications for the conduct of financial intermediation and external investment funding. As regulatory considerations recede, the premium placed on the economic rationale of securitisation occurs at a time when the origination of loans has become a fiercely contested business. In the quest of more efficient banking operations banks are pressed for enhanced credit risk management capabilities and allocative efficiency in loan origination. Both aspects underpin the economic rationale ofsecuritisation under an internal ratings based regulatory framework. Given the competitive nature of capital markets, improved risk-adjusted returns are likely to translate into more favourable loan terms for bank debtors that qualify for standardised credit assessment and wish to partake of standard loan contracts with minimised idiosyncratic risk. The dependence of profitable asset securitisation on the acquisition of off the shelf loans does inevitably bias financial institutions into altering the composition of their loan book for purposes of cost efficient asset funding. The illiquid nature of customised loan contracts coupled with higher information cost, non-standardisation will carry a premium compared to standardised credits, even if the risk involved is the same. Investment funding, such as project finance and SME finance, is becoming less attractive to banks and non-bank financial institutions as the information of private information in a close borrower-lender relationship or the entrenchment of individualised service defies accurate pricing in securitisation markets. In pursuit of cost efficiencies banks would for the most part be inclined to forgo customisation, as the ease of subsequent securitisation drives the acquisition of debtors, i.e. the degree of standardisation of assets determines the cost of securitisation. Hence, non-standard loans will remain to be offered, but only at a higher price (which might increase adverse selection and credit rationing).

What appears to be turning the principle of traditional bank-based financial intermediation upside down, is nothing other than a re-definition and fine-tuning of the intermediation process. Like in traditional deposit business, the terms of the lending business under securitisation is conditioned on the cost of asset funding and its attendant exposure, that is, the cost of capital sets the reference base for adequate contribution margins in asset origination. This interpretation of asset funding in securitisation preserves the concept of financial intermediation, with the exception that securitisation effectively disintermediates deposit-financed bank credit (deposit business). Issuers of securitisation transactions subordinate investor claims by connecting investors of various risk appetites directly with debt securities structured to meet commensurate risk tolerance of investors. Asset funding through the origination of debt securities forges a new process of intermediation, with the deposit business taking a backseat. Nonetheless, loan securitisation--with banks acting as loan brokers capitalising on their informational rents--continues to be grounded in the idea of banks as conduits of efficient allocation of investment funds. Loan securitisation modifies the criteria of lending business and advances an efficient asset funding process, defined by how far the loan book can be restructured to meet the demands of issuing structured claims on an underlying loan portfolio. In other words, the diversification effect and the reduction of economic capital in securitisation is proportional to the use of standardised bank loans, once regulatory arbitrage has been rendered less profitable.

From a regulatory point of view, bank-based loan securitisation might display the same characteristics of credit risk as traditional lending, depending on the payment structure and the security design of the securitisation transaction. As banks tend to retain a significant portion of credit risk in the form of credit enhancement in combination with complementary structural enhancements, CLOs pose prudential issues of credit risk management similar in scope and significance to conventional lending business. However, the elaborate security design of loan securitisation commands a regulatory treatment of credit risk of structured finance more comprehensive than what is currently considered in banking supervision of traditional lending business. Apart from issues of financial stability, improved credit risk management techniques applied by issuers of loan securitisation transactions attributes greater significance to aspects of investor protection. From a regulatory perspective, such structured finance investments may need extra supervision to reduce threats to the global financial system, as the inherent complexity should not blind the beholder to the fact that unregulated financial institutions pose a threat to the stability of financial markets worldwide, unless rules on risk management, transparency and investor protection prove to be adequate (Eichel, 2002).

6. Implications of the regulatory system and other general trends on bank lending--securitisation as a conduit of regulatory constraints

Future changes in the conduct of credit risk management and the lending policy of banks are not so much driven by the requirements for securitisation in the pursuit of lower economic mad regulatory capital, but rather by the radical change in mutually reinforcing trends challenging banks to be more efficient in the management of credit business. That is, securitisation epitomises one possible vehicle of such efficient change management.

First and foremost, the fundamental shift in the regulatory system governing financial intermediation is one trend that has induced a changed business paradigm. Devised as a arbitrage mechanism to exploit regulatory shortcomings, securitisation is no longer limited in application to opportunities arising from the one-size-fits-all treatment of credit risk in the current regulatory system, but also caters to the anticipated regulatory change as regards less standardised procedures applied in the determination of capital requirements for credit risk exposures (Basle II). Irrespective of the approach chosen for the calculation of capital adequacy (Standard Approach, Foundation Internal-Ratings-Based Approach (Foundation IRB), Advanced Internal-Ratings-Based Approach (Advanced IRB), the implementation of the new proposal of the Basle Accord, in one way or the other, requires a re-definition of banking operations in order to increase the liquidity of loans. The use of securitisation and credit derivatives makes a good subtext to this change process induced by new regulatory reform. The following the core aspects of reform in bank lending are particularly amenable to securitisation:

(i) a consistent internal rating and scoring model on an individual debtor basis, (ii) a detailed calculation of individual risk exposure in order to establish a transparent creditor-debtor relationship, and (iii) comprehensive and active credit portfolio management for purposes of avoiding risk concentrations (granularity), which might serve as a basis for the implementation of risk control routines.

Under (ii) increased levels of sophistication in credit risk assessment allows for an accurate identification of concentrations of risk exposures as percentage of economic capital in excess of a certain absolute risk tolerance (see Exhibit 24). Diversifying these risks would require a careful consideration of both concentration and correlation effects of individual exposures contingent on a given portfolio size, i.e. the exposure-weighted number of assets.

[GRAPHIC OMITTED]

Exhibit 25 above illustrates aspect (iii)--the relationship between sccuritisation as an operational response of financial institutions to turn the tide of declining yields from interest-based business, on the one hand, and regulatory reforms set forth in the new proposal of the Basle Accord and active credit portfolio management, on the other hand. Hence, any active management of credit risk involves a securitisation process, which determines the reference base for a risk-adjusted lending policy.

[ILLUSTRATION OMITTED]

Secondly, the prospect of under-performing credit assets as well as a legacy of poor pricing and cross-selling in interest-based business, such as lending to corporates and sovereigns, has led banks to embrace securitisation as a convenient tool to overcome regulatory and economic capital constraints. Apart from such internal demand-driven reasons of changes in the way banks manage lending business and attendant credit risk, the pervasiveness of methodological advances in credit risk assessment (see Exhibit 26 above) and sophisticated portfolio analytics have helped establish structured finance transactions as an essential refinancing tool of banks and financial service firms (Oliver, Wyman & Co., 2002).

[ILLUSTRATION OMITTED]

Consequently, the interaction of these trends (see Exhibit 26) emphasises the critical importance of active credit portfolio management, sustained by consistent high-quality credit risk analysis that follows a tried and tested methodology closely aligned with one or even more current credit risk modelling techniques. First, endogenous credit risk models break down into two categories: the credit migration approach, which is based on the probability of credit quality moving from one rating classification to another (transition probability), including default, within a given time horizon (applied by JP Morgan with CreditMetrics) and the option pricing approach or structural approach, which rests on the asset value model originally proposed by Merton in 1974, where the risk exposure of the capital structure of a given firm follows an endogenous default process. Default occurs when the value of the firm's assets falls below some critical level. Second, the actuarial approach applied by Credit Suisse Financial Products (CSFP) with CreditRisk+ only focuses on default for individual bonds or loans, which is assumed to follow an exogenous Poisson process. Finally, the econometric approach proposed by McKinsey in CreditPortfolioView follows a discrete time multiperiod model where default probabilities are conditional on macroeconomic variables.

Yet, it is anybody's guess if banks are willing to realign business roles and responsibilities in a lending process as illustrated in Exhibit 28, which takes account of both strategic changes in the lending business and the importance of credit risk assessment. The implementation of these core aspects of active credit portfolio management lead to a radical redesign of business processes in bank lending, provided that risk control routines take into account credit volume. As the origination of loans and portfolio investment is unbundled, the risk-oriented determination of credit conditions and increased efficiency in the lending process through standardised credit terms are essential components of a new organisational model of bank lending. Therefore, securitisation of loans and other bank assets would lead to a flexible structuring of the credit portfolio if market prerequisites are satisfied.

[ILLUSTRATION OMITTED]

The strategic consequences for the lending policy of banks will vary between large and small banks. Large banks, with more sophisticated credit management systems, are better prepared for an internal-ratings based determination of minimum capital requirements, which lead to a more truthful representation of the risk-return relationship in the lending business. With loan pricing under the new regulatory framework geared to internal credit risk assessment, large banks will be ill at ease to fully transpose exposure to poorly rated loans into higher risk-adjusted spreads. The perpetuation of past standards of loan origination appears hardly feasible. Since the attractiveness of a loan is also continent upon both the credit portfolio quality of the loan book and the corresponding routines for credit risk control, large banks are more inclined to focus on the strategic business of highly rated loans (The Boston Consulting Group, 2001).

Smaller banks are faced with the need to adjust loan terms in accordance with capital requirements of the standard approach in the Basle II proposal to preserve their competitiveness; however, lower flexibility in the calculation of capital adequacy (e.g. incorporation of loss given default (LGD) in a standardised calculation of credit risk exposure, etc.) defaults smaller banks into accepting those levels of credit risk, which attract more beneficial treatment in the standard approach of credit risk weighting--namely poorly rated loans with high interest margins. Moreover, even if bad risks were to be weeded out, the process would not offset the cost of restructuring the credit portfolio. Although the standard approach enables smaller banks to comply with regulatory minima by adopting a risk-adjusted lending policy in line with more advanced portfolio credit risk management, low quality debtors with long maturity loans are most likely to migrate to smaller banks, creating larger credit risk exposure. (48)

Consequently, the administration of securitisation by financial intermediaries is the consequence of a more responsive regulatory system and new external constraints, which reward increased sophistication of internal credit risk management--be it driven by either efficiency gains or regulatory incentives or both. Banks would focus on underwriting, product engineering, distribution and trading of structured finance products through the active use of credit derivatives in order to achieve favourable tax and regulatory treatment of their loan portfolio. Nonetheless, securitisation is only one way to address more sophisticated credit risk management. Besides a securitisation model, banks could also adopt other operational structures in anticipation of future business end games for financial intermediaries (see Exhibit 29).

Exhibit 29. Models of possible "business end games" for
financial intermediaries in the wake of regulatory change
(Oliver, Wyman & Co., 2002)

"INVESTMENT BANKING MODEL"

* banks acting as intermediaries

* activities focused on loan underwriting, distribution,
secondary trading

* end-investors: mutual funds, insurance companies, asset
managers

* favourable tax and capital treatment

"REINSURANCE MODEL"

* parallel to P&C (property and casualty insurance) market

* banks hold specific risks of individual loans

* reinsure against "tall" risks (large losses)

* insurance opportunity for capital rich, globally-diversified
institutions

"SECURITISATION MODEL"

* ultimate manifestation of increasing liquidity

* banks focused on underwriting, product engineering,
distribution, and trading

* active use of credit derivatives

* favourable tax and capital treatment

In an investment banking model banks would specialise in their role as intermediaries for end-investors, such as mutual funds, insurance companies and asset managers, with their core activities limited to loan underwriting, distribution and secondary trading for purposes of limited tax expense and capital cost. Finally, if the internal risk management routine of a bank reaches a level of sophistication sufficiently advanced, such that a bank could profitably accept specific risks of individual loans, a reinsurance model becomes feasible. In this model of a business end game banks focus on reinsuring counterparties against large credit risk exposures--parallel to the property and casualty insurance market, likely to be limited to capital rich and globally diversified companies.

CONCLUSION

In a nutshell, the paper explained conceptual issues arising from the economic rationale of loan securitisation--in the form of collateralised loan obligations (CLOs)-as an efficient refinancing mechanism, mainly motivated by the mitigation of economic and regulatory cost of capital, balance sheet restructuring and hedging incentives. After a general classification of asset-backed securitisation (ABS), we set out to illustrate the typical security design and the degree of information economics (adverse selection and moral hazard) loan securitisation involves. Subsequently, we shed light on the analytical consequences arising from specific characteristics of securitisation, such as the definition of structured ratings and the function of credit enhancement. This exposition of the various components of securitisation guide our understanding of the interactive effects between securitisation, on the one hand, and loan portfolio composition, credit risk exposure, asset funding of banks and bank regulation, on the other hand. Judging by the latest developments in banking finance the feasibility of loan securitisation might have become a child of its own making due to a responsive regulatory system and rising sophistication of active credit risk management.

Exhibit 21. The balance sheet of a special purpose vehicle
(SPV) in collateralised loan obligations (CLOs)

The SPV fulfills the following conditions:

* protected from insolvency of the originator
(bankruptcy remoteness)

* no recourse to the originator (non-recourse
clause)

* must not fall under corporate taxation
because double taxation would make
transaction too expensive

* usually organised as a trust and not
consolidated with the originator

* Its business activities are limited to the
issuance of predefined securitisation
structure

* further indebtedness is not possible

Balance-sheet of a SPV in an unlevered CLO transaction
(in U.S. $m)

  Assets    Liabilities      Rating       %
(credits)     (notes)

  5.000        4.650         AAA/AA       93
    --          100            A           2
    --          100           BBB          2
    --          150#          N/A          3
    --
  5.000        5.000       Transaction    100
                             Rating

# (unrated) equity tranche retained by sponsoring entity

In a SPV-based structure of securitisation the risk of the
securitised assets is totally separated from the originator.
For the investors, only the SPV is the liable party

Exhibit 27. The three basic approaches to credit
risk (portfolio) modelling

               Endogenous Models             Econometric Models

             (credit migration approach
               and structural/option
                 pricing approach)

Risk and       * estimated default           * estimated
Correlation      frequency (EDF)               default
Parameters       asset correlations            frequency (EDF)

               * industry/country            * sensitivities to
                 weights                       macroeconomic
                                               variables
               * (loss severity)
                                             * (loss severity)

Basic          * firm value follows          * default rate
Analytics        default process. Credit       is regressed
                 event occurs if value         on macroeconomic
                 of assets is less than        variables and
                 value of liabilities          random
                 [Merton-based/structural      innovations
                 approach]
                                             * Monte Carlo
               * default probability           simulation
                 depends on the                of default rates
                 probability of credit         generates loss
                 risk migration to             distribution
                 reach the default state
                 [credit migration
                 approach]

               * Monte Carlo
                 simulation of correlated
                 asset returns and default

               * generates loss
                 and/or NPV distribution

Commercial     * JP Morgan CreditMetrics     * McKinsey
Applications                                   Credit Portfolio View
               * KMV Portfolio Manager
                                             * proprietary
                                               portfolio models

                Actuarial Models

Risk and       * estimated
Correlation      default
Parameters       frequency (EDF)

               * default rate
                 volatility

               * sector weights

               * (loss severity)

Basic          * default rate is assumed
Analytics        to be a random variable

               * closed form solution
                 for default and loss
                 distribution

               * generates loss
                 distribution

Commercial     * CSFP CreditRisk+
Applications

Andreas A. Jobst

London School of Economics and Political Science (LSE) and J.W. Goethe Universitat Frankfurt am Main