IV. Analytical consequences arising from the characteristics of securitisation.

Despite the aforementioned benefits associated with banks engaging themselves in loan sccuritisation, securitisation activities warrant the application of prudent conduct and due diligence. Otherwise, they could increase the overall risk profile of the issuer. Even though the securitisation of loans involves the same degree of risk exposures as bank lending business, which stems from credit risk, interest rate risk (including risk from prepayment), concentration risk, operational risk and liquidity risk, the Basle Committee confirms the notion that unbundling the traditional lending function into several limited roles is prone to inflict more complex credit risk on issuers of CLOs. With the credit risk being shared between several stakeholders in the securitisation process, i.e. originator, servicer, sponsor, credit enhancer, liquidity provider, underwriter trustee, investors and, as need be, credit derivative counterparty, an additional layer of administrative and processual complexity confounds the aforementioned types of risk, whose impact on the risk sensitivity of the reference portfolio might be markedly different from traditional lending. This observation especially applies to interest risk, since securitised loans are commonly regarded more sensitive to interest rate movements than unsecuritised loans, as they display a higher positive correlation between the probability of rating change to interest change. However, by way of explaining the different nature of risk in structured finance transactions, the consequences for the assessment of investors' claims on the reference portfolio of a CLO are straightforward. Whereas the quality of the reference portfolio (credit rating) is determined by the assumptions entering the credit risk assessment and management system of the issuer, structured ratings have to take into account the complex nature of CLOs and the diverse risk patterns imposed by the various agents in the securitisation process.

A. Credit Risk--An Issue Of Diversifiability

We first need to shed light on what actually constitutes credit risk and how the properties of credit risk warrant particular methods of mitigating its adverse effects on the bank loan book. From a management perspective, Oldfield and Santomero (1997) argue that uncertainties, i.e. risk associated with the completion of conditional counterparty promises, facing all financial institutions can be segmented into three separable types. These are:

* risks that can be eliminated or avoided by simple business practices,

* risks that can be transferred to other participants, and

* risks that must be actively managed at the firm level.

Credit risk arises from non-performance by a borrower, which is caused by either an inability or an unwillingness to perform in the pro-committed contracted manner. If two parties engage in a loan contract, i.e. a contractual obligation such that funds are transferred from one party to the other for an agreed period of time in return for compensation in the form of interest, the probability of commitments to be honoured gives rise to such inherent uncertainty.

This uncertainty can affect the lender holding the loan contract, as well as other lenders to the creditor. The risk inherent in the intertemporal compensation for the periodic transfer of wealth between the provider and the recipient of capital in external finance warrants an accurate prediction as to its default probability and loss severity. Such as measurement is irrespective of the means of external finance, be it corporate bonds, bank loans or any other form of debt securities or credit obligations, such as underfunded pension provisions. Therefore, the financial condition of the borrower as well as the current value of any underlying collateral is of considerable interest to its bank. In order to prevent the assessment of financial strength as a proxy of the probability of repayment at the agreed terms and conditions set out in the loan contract from being of conjectural nature, banks employ clearly defined credit rating systems in order to quantify the capacity of lenders in generating sufficient future cash flows to meet scheduled repayments required by the lender. Credit ratings address the likelihood of full and timely payment of principal and interest to lenders and expenses of other third parties involved.

Given a portfolio of loans, the real risk from credit lenders face is the deviation of portfolio performance from its expected value. Accordingly, once standardised and made comparable, credit risk is diversifiable, but difficult to eliminate completely, as market risk and interest rate risk as systematic impediments defy diversification to the effect the financial strength of debtors and the funding sensitivity of the creditor respectively. With respect to the transferability of credits, loans cannot be distinguished along the lines of diversifiability of risk on the basis of an exclusive distinction of either interest rate risk exposure or idiosyncratic credit risk respectively. While large part of credit risk can be diversified through the optimisation of the "diversity score" of loan portfolios according to portfolio theory, an interest rate risk portion inherent in a loan portfolio poses a systemic risk that can only diversified to a limited extent. Banks rather employ hedging technique to mitigate possible default losses in this respect. However, the degree of this remedial mitigation of risk impediments is bounded by the transferability of systematic risk of assets. Furthermore, the idiosyncratic nature of some portion of loan losses remains a problem for creditors in spite of the beneficial effect of diversification on total uncertainty. This is particularly true for banks that lend in local markets and ones that take on highly illiquid assets and agree to unsecured repayment schemes (e.g. non-mortgage, non-collateralised industry loans, etc.). In such cases, the credit risk is not easily transferred, and accurate estimates of loss are difficult to obtain due to the residual specific risks in the loan book.

B. Downgrade Risk (Credit Quality) And Claims-Paying Ability--Fundamental Components Of Credit And Structured Rating

1. Definitions

The discrepancy of credit ratings and structured ratings in loan securitisation stems from the different assumptions, which enter the estimation of credit risk associated with an outstanding payment obligation. By common consent, rating agencies distinguish two types of assessment methodologies of a debtor's credit posture, the analysis of downgrade risk and the analysis of claims-paying ability. Even though the determinants of downgrade risk and claims-paying ability are closely related, the two concepts exhibit areas of analytical distinction. For descriptive purposes we consider an abstract, albeit overly simplistic, definition of both concepts similar to Canor et al. (2000). Whereas the claims-paying ability speaks to the probability of debtor default on some obligation, the downgrade risk reflects the probability that modest changes in the financial condition of the obligor might induce a reassessment of its claims-paying ability in the Suture. This "benefit of doubt" in the future development of credit quality is akin to the underlying principles governing the forward-looking rationale of conventional credit rating, which focuses primarily on the question of whether the financial resources at the disposal of an obligor match up to the stochastic loss severity incurred in the event of adverse shocks (determination of risk exposure), irrespective of a change in the macroeconomic environment.

Since the claims-paying ability is an inherent component of the assessment of downgrade risk, the same factors that help obligors to maintain solvency tends to limit their susceptibility to rating downgrade in the long run. In estimating claims-paying ability, however, the accumulation of a large number of adverse developments is assumed to be the most likely scenario to precipitate defaults. This includes the sharp deterioration of their macroeconomic environment. So the risk estimate of the claims-paying ability describes the obligor's expected ability to sustain long-term distress, i.e the degree of financial strength in the wake of dramatic portfolio deterioration. Such a present reflection of a worst-case scenario comes to bear in cash flow projections and adequate provisions of risk cover in structured finance transactions. In the move likely to support structural resilience to portfolio credit default in stress scenarios, several factors limit downgrade risk in the long run (Canor et al., 2000): (i) low-risk and highly diversified reference portfolios of loan risk, (ii) high levels of transparency, (iii) limited non-financial guaranty activities, (iv) robust and consistent capital bases, and (v) prudent exposures to market risk in general and liquidity risk in particular.

The intertwined causality of claims-paying ability and downgrade risk across time occasions compelling plausibility for a strong credit posture of businesses. Maintaining the credit rating at a comfortable level is vital to a business franchise in order to avoid being caught off" guard in times of prolonged capital shortage. For instance, only if the obligor is deemed to "stay within the fairly circumscribed financial guarantee business model" (Canor et al., 2000, 3) in the context of structured finance, the current rating level will not be subject to downgrade risk. In cases of significant capital demand in anticipation of strategic changes and significant investment, obligors will take sustained efforts to avert downgrade risk in preserving their financial capabilities.

Since the downgrade risk indicates the risk profile of the obligor over time, the chance of credit risk deterioration and other adverse changes in financial fundamentals might warrant a downward rating adjustment. A variety of reasons could put pressure on the rating of obligors. Alterations in the business strategy and the competitive situation could lead to damaging unexpected credit deterioration or large unexpected losses in the underlying reference portfolio. In the case of structured finance transactions, rating downgrade might also be induced a great deal by the quality of underwriting and creditor surveillance of the issuing party. Moreover, the sponsor's long-term capital plan in funding the reference portfolio could be inconsistent with the projected payments on the individual assets included therein, given the assigned rating of both the portfolio and the guaranteeing issuer. In anticipation of weakened financial strength of obligors, downgrade risk focuses on the comparison of financial fundamentals (credit posture) and the probabilistic impact of adverse shocks independent of the macroeconomic environment. The significance of rating change is indistinct, however, unless one is able to identify a good leading indicator of altered downgrade risk.

Measuring the financial condition of the obligor's core franchise (in the case of credit rating) or the sustainability of projected cash flows from a reference portfolio of a securitisation transaction (structured rating) could serve to this end. Whenever one observes a strong business model with its attendant credit strength, the obligor can be safely assumed to generally have both the incentive and the resources to remedy capital shortages arising from unexpected credit deterioration. To the contrary, a weak core business could induce obligors to consider alternative, more profitable, though riskier, areas of operation. Such operational change might coincide with a capital structure that inhibits flexibility of raising additional funds and would force the obligor to exercise restraint in earnings dilution. The same rationale applies to structured ratings in securitisation transactions. Any so-called unscheduled variation (i. e. unexpected losses) in portfolio credit quality is addressed through either early amortisation or structural enhancements to avert downgrade risk, which equates to a long-term absence of claims-paying ability. However, if excessive credit deterioration renders the reference portfolio incapable of satisfying contingent claims of investors and erodes the claims-paying ability), capital needs are poised to effect increased downgrade risk.

2. The importance of downgrade risk (credit quality) and claims-paying ability in loan securitisation

The common security design in securitisation has it that the rating of the reference portfolio and its guarantor reflects investor sentiment about the likelihood of expected returns over the lifetime of the transaction. The credit risk assessment of loan securitisation generally involves developing a portfolio risk model, which describes the probability distribution of potential credit losses of the reference portfolio. In securitisation we concentrate on the claims-paying ability of the guarantor in judging the quality of the reference portfolio. So, a structured rating reflects a comparison of available capital resources, i.e. the value of the underlying loan pool and the liquidity provision, and the loss severity of credit deterioration in extreme distress scenarios. Given its sensitivity to systematic risk of macroeconomic change and endogenous constraints of counterparty risk in loan contracting (average portfolio credit quality and risk concentration), the claims-paying ability is indicative of credit rating development over time. Since the sustainability of a rating is rationed on the grounds of its resilience to stress scenarios, the issuer's reference portfolio and the security design of a securitisation transaction includes substantial credit risk provisions. So the present stability of capital resources translates into reduced downgrade risk over the lifetime of the transaction.

Despite the intertwined determinants of both downgrade risk and claims-paying ability, the fundamental determinants of credit risk in loan securitisation--credit risk concentration and average portfolio rating of the reference pool of loans--might be less relevant to the comparison of the current financial condition and projected worst-case loss severity (in testing the claims-paying ability). They appear to have greater effect on the probability of rating deterioration implied in downgrade risk rather than the ability of surviving financial distress. Hence, the severity of downgrade risk to portfolio concentration corresponds to its sensitivity to macroeconomic change. The probability of an obligor to stomach a general economic downturn arguably increases in the level of diversification and the degree of portfolio quality. Although idiosyncratic risk is permanently mitigated through prudent portfolio choice, the distinction between certain portfolio risk concentrations and their attendant credit exposures is likely to be more pronounced in times of normal economic development (anti-cyclical differentiation). In contrast, the adverse impact of a recession would hardly discriminate against differently rated portfolios and could subject obligors to what could be considered a sweeping trend towards generally higher downgrade risk in disregard of different ratings. In absence of boom or recession, however, credit risk assessment is much more suited for a fine-tuned differentiation between different rating grades over long-term stability rather than short-term stress tests.

C. The Function of Structured Ratings

It is worthwhile bearing mind that expectations about the support mechanism of a securitisation transaction are vital in interpreting the rationale of corporate ratings and structured product ratings respectively. Certainly, ratings for structured finance products rest to a much higher degree on quantitative parameters than public ratings on corporate bonds or public debt, which tend to incorporate all information about an entity that is known at present. However, in credit ratings the issuer quality encapsulates contingent adverse events only to the extent that they affect a revision of current rating, as determinants of corporate performance and/or the business environment will change in the future. This limited state-contingent perspective does not project probable future trends but qualifies as a measure of current creditworthiness and financial performance. In some sense corporate ratings are upfront estimates without stochastic contemplation of future outcomes.

This benefit of doubt is not apparent in the context of structured product ratings. Here, the current rating reflects the present value of a worst-case scenario applied to the expected future cash flow generation of the reference portfolio, based on the assumed occurrence of a predetermined trigger event. Given a certain credit risk rating at present the transaction is to survive a multiple of rating-based default loss, such as stress testing, which exposes asset performance to adverse scenarios exceeding historical norms in terms of interest rate mismatch, payment rates and recoveries. The process of determining the credit enhancement has to take into account not only today's risk but also future risk factors that profoundly impact the value for the transaction. Essentially, the incorporation of future unknowns elevates structured ratings to apply stochastic prudence in estimating asset performance, given the security design of the transaction. This includes vital components of a transaction. Hence, commensurate credit enhancement in a structured finance transaction has to be inadvertently conditioned on the present projection of future outcomes, albeit the side-effect of a marked reduction of rating volatility as opposed to corporate ratings. In the event of under- or over-performance, i.e. the deviation of actual observations from expectations, a change in structured ratings ensues. Consequently, one is sure to expect structured ratings to display lower volatility than credit ratings.

The difference of corporate and structured ratings is also apparent with respect to the structural properties underlying the asset in question. Given the formal rescinding of credit-linkage, the rating of securitisation transactions aims to effectively disentangle credit and transaction/structural risk in assessing the quality of both the underlying reference portfolio and the corporate rating of the sponsoring agent/ issuer. Consequently, structured ratings have to honour the distance of default between these two rating reference cases. Thus, only third-party support incorporated in the structural features of the transaction, such as the credit enhancement, should come to bear in the determination of the rating associated with structured ratings. In contrast, the corporate ratings comprehensively assess the default risk and propensity of adverse business development of one entity only, irrespective of supportive mechanisms in place, which might distort and bias the results. The estimate of financial strength frequently determined in the context of public banks does reflect this distinction most prominently.

Since any turbulence in credit markets profoundly impacts on the performance of structured finance transactions--such as collateralised loan obligations as a subset of collateralised debt obligations (CDO)--ex ante (i.e. presale) rating of structured assets is closely coupled with need of proactive surveillance of future rating performance in terms of loan quality and possible recovery rates is critical in averting consistent deterioration of the par value of the reference portfolio securing the structured transaction. Any erosion of portfolio value takes a heavy toll on the credit support mechanism especially in the case of true sales rather than synthetic on-balance sheet securitisation structures. The loss absorbing capacity of credit enhancement supporting debt securities issued by the sponsor or the special purpose vehicle in traditional securitisation depends on the overall performance of the reference portfolio rather than a nominal amount of third-party credit risk coverage. A persistent increase in expected loss rates of the loan pool and a steady downward rating migration signals weakening collateral performance, which inflicts negative bias on the credit quality of performing assets within the reference portfolio. Eventually, this development further exacerbates the maintenance of adequate credit enhancement in order to shield the transaction against adverse collateral performance

D. The Management of Risk From The Issuers Perspective

As banks and non-bank financial institutions engage in ever more complicated structures of loan securitisation, the credit risk flowing from such transactions has necessitated commensurate methods of assessment and control thereof. In the light of increased sophistication in structured finance, statistical procedures, which were originally developed for traditional banking business, such as on-balance-sheet lending or plain vanilla transactions in debt securities, have been refined to satisfy the rising demands in credit risk management. Financial institutions address this issue by means of comprehensive credit scoring systems, which are based on statistical procedures that provide an estimate of default probability and loss severity for loans selected for securitisatlon. Although the standardising nature of credit scoring has been particularly amenable to known obligors with sufficiently high transparency, the scoring methodology is being increasingly used for small business lending and middle-market commercial lending.

As credit scoring contributes to consistency in credit origination and loan underwriting standards, it has become intimately tied with loan securitisation. By the same token, a more accurate and comparable estimation of the loss probability distribution for a loan pool being securitised is highly desirable. The lower the variance in predictions about estimated default losses arising from a selection of loans, the lower are the efficiency losses in the external assessment of asset quality by rating agencies. Credit scoring models require historical data of credit events on a large sample of fairly homogenous loan contracts in order to derive good estimates of expected losses. The standardisation of loan agreements arguably supports the process of generating such scientifically meaningful and reliable data.

Being at the intersection of internal credit risk management and public scrutiny of asset quality in capital markets, securitisation heavily relies on accurate estimates of loss given default of loan reference portfolios. Needless to say, financial institutions with proper in-house credit risk management capabilities consider securitisation an attractive method of refinancing if credit scoring models can be used in the origination process of standardised loan agreements. Thus, combining credit scoring models with a standardised lending policy reduces non-interest expenses associated with lending from the perspective of issuers of securitisation transactions.

Increased efficiency and standardisation also has crucial implications for the cost of securitisation contingent on the ease of external credit risk assessment by rating agencies. For example, the amount of credit enhancement--the degree to which the issuer of a loan securitisation provides credit risk cover for first losses of the reference portfolio--is a key cost driver in such transactions. As the use of credit scoring and loan standardisation improves the statistical power of credit risk estimates, credit rating agencies are better able to determine how much credit enhancement is needed for issuers to achieve a desired rating on a given reference portfolio (with a certain average default probability). Consequently, a bank can eliminate a substantial portion of this underwriting cost in loan origination when competing for new business.

Andreas A. Jobst

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