Executive Summary

In the last December, 2014, the Basel Committee issued a new consultative paper, for a relevant revision of the credit risk “standard” model. This is just a piece of a global strategy for the revision of the standard models. In fact also the market risk and the counterparty risk, with some recent ongoing processes, are deeply changing in the standard calculation mechanisms. The paper shortly reviews the current situation and points out the main innovations, along with some issues that must be better defined.

1 The Current Credit risk capital requirement taxonomy 

The Basel III framework, started on 2014, January 1st, focused mainly on:

• New capital charges building-blocks, such as the CVA for the counterparty risk and the systemic risk for the SIFIs banks
• New constraint classes, such as the liquidity and the leverage indicators.

The “classical” credit risk of the banking book was not changed in an effective way. Among the EU (CRDIV) specific updates to the general framework, we recall the “SME supporting factor”, that allows to apply a smaller risk weight coefficient to the counterparties where the revenues are less than  1.5 mln €.

Hence also the current regulatory set-up for the credit risk may be referred to the Basel II rules.

We recall the capital requirement calculation strategies.

The capital requirement may be calculated according to two different approaches, the standard approach (we observe that in the Basel II naming often the term “standard” is used, while “standardised” is preferred in the new revisions) and the internal rating approach (IRB). More explicitly:

• In the standardized approach the capital requirement is given by K = EAD × CCF × RW × 8%, where CCF = credit conversion factor and RW = risk weight. In the most of cases CCF = 100%. We have some differences e,g. in the personal facilities where CCF = 0.5. RW strongly depends form the rating of the counterparties and from the reliability of the rating agency. The RW may be 0%, 20%, 50% … 1250% as a grid depending form the Basel portfolio (sovereign, banks, corporates, retails, high risk exposures) to which the counterparty/transaction belong and from the rating agency. Some specific encouraging coefficients are given for the mortgages, for the residential real estate we have (CCF × RW) = 35%, while for the commercial real estate it grows to 50%. Finally, if the bank does not adopt any rating agency and/or for the unrated counterparties, RW = 100%.
• In the IRB approach one has K = EAD × LGD × f (PD,ρ,M). PD is the default probability of the debtor, M the maturity of the exposure (with some cap and floor level, not actually used for all exposures), r the correlation coefficient assigned by the Basel committee to each portfolio. The function f( ) is well known as the risk weighting function and it wants to mimic a 99.9% 1 year CreditVaR under some simple assumptions. See [7]. More specifically, we have the basic approach, where the bank can estimate only the PD coefficient (LGD and EAD being assigned by heuristic rules), and the advanced approach, AIRB, where the bank can develop its own models also for the exposure and the loss given default values.

2 The Basel Accord and the rationale for the new standardised models

The consultative paper for the new  credit risk standardised model follows and comes along with the previous relevant revisions of the market risk (paper 265) and counterparty risk (paper 279).

This general evolution comes from some evident drawbacks of the Basel 2.5 and Basel 3 reforms, namely:

• The internal models with these reforms implied some double counting effects, such as in the Stressed VaR measure and in the IRC capital charge.
• On the other hand, the standardized model did not have significant increases in the capital requirement, except than some update in the coefficient for the specific mark risk.

See [1], [4], [5] and [6] for a complete review.

Due to the above reasons in the market one can observe an evident paradox, i.e. the capital requirement for the banks that have invested a lot for the internal model validation often shows figures higher than the simple standard models.

Hence the Basel committee is trying to arrange this inconsistency. Moreover, the general goal is that the standardised new models must be more risk sensitive (i.e. correlated to the effective risk drivers) and more granular, i.e. more flexible in being adapted to the specific bank exposures/portfolios.

These revisions should satisfy (at least) two different technical issues: 1) a tradeoff between accuracy and simplicity. The small banks have not the budgets or the know-how for too sophisticated approaches. 2) a robust calibration rules set, in order to level the playfield (with respect to the internal models) and to update the coefficients when needed.

3 The new standardised credit risk model

3.1 The general principles

As stated in the paper 305, the review want to math several purposes, that is:

• reconsider the design of the standardised approach to ensure its continued suitability for calculating the capital requirements for credit risk exposures;
• ensure that the standardised approach is appropriately calibrated to reflect to a reasonable extent the riskiness of exposures;
• increase comparability of capital requirements under the standardised approach and  the internal ratings-based (IRB) approach by aligning definitions and taxonomy, where possible;
• increase comparability of capital requirements between banks using the standardised approach by reducing national discretions, where feasible; and
• reduce reliance on external credit assessments by providing alternative measures for risk assessment, where possible.

The above purposes are stated in some more detailed seven principles. The reliance from the external ratings is one of the most innovative principles. At a first level, one can doubt about its applicability, because of the current “propagation” of the (external) rating logics into the whole financial market participants.

Generally speaking, all the new proposals aim to define in a clear way the risk drivers underlying the exposures. The availability of practical guidelines about the risk driver makes it possible to satisfy the risk sensitivity approach.

3.2 Some details

Let us classify the main new proposals according to the different exposures categories:

Exposures to the Banks

Currently the standard model gives two options. 1) to apply the risk weight depending from the sovereign rating of the country of the bank 2) to apply the risk weight related to the bank rating.

The new proposal suggest as risk drivers the following:

• the bank capital adequacy, e.g. as measured by the CET1, common equity tier1. Another choice could be the Tier1 or the leverage ratio
• the asset quality of the bank. The asset quality might be considered a good default predictor. The indicator is called NPA

Finally, some rules to include the “securities firms” in the bank exposures portfolio are given.

Exposures to the Corporates

As a main risk drivers, the committee indicates the  revenues of the firms.

The rationale is that any profitability measure could create some misaligned incentives and introduce excessive procyclicality.

The second risk driver is the balance sheet leverage.

The risk weight could range from 60% to 300% of a suitable combination of the risk drivers.

Retail Portfolio

The current standard model assigns a 75% preferential coefficient to the retail exposures. The consultative paper enhances the criteria to qualify for the preferential treatment in the retail category.

These criteria can be sketched in the following points:

1. Orientation criterion;

2.  Product criterion;

3.  Low value of individual exposure criterion

4.  Granularity criterion.

The committee will evaluate if maintain the unique 75% risk weight or to apply a diversified table, as a function of concentration and maturity.

3.3 Issues and Criticism

Many banks and banking association have given feedbacks to the consultative papers. We cannot give an exhaustive review, but let us summarize as an example the perspectives of some very different players:

• The French Banking Association. The main skepticism is related to the parameters of the calibration that could increase the capital charge for bank exposures in the 20-30% range and the corporate exposures in the 40-60% range. Moreover, a set of multiple QIS is asked, as the calibration can not be point-in-time. Furthermore the external ratings are suggested not to be removed, since they can be used as a complementary risk driver for the banks and large corporate exposures.
• The Italian Federation of Co-operative Credit Banks (FederCasse). The main remark is related to the possible increase in the capital charge. This is motivated by the proposed factors and parameters and by the credit portfolio features of the small banks, i.e SME s and retail portfolios. A “local” calibration process is suggested, within a set of well stated rules. Moreover, the use of external ratings should not to be removed, as they have shown an acceptable reliability. Finally, a proper time horizon for the QIS and phase-in period is claimed.
• Deutsche Bank. The arguments are very similar to the previous ones. More time is requested for the QIS and calibration exercise. DB doubts that a simple 2-risk drivers model can be more sensitive than the external ratings. Hence the request is not to remove their usage. Finally it is pointed out that the new capital charge need an extensive set of data, that typically is not available for the banks without internal model. The data lack could imply a too punitive risk weight, e.g. 300%.

The debate in the banking sector is still in progress, and the timeline for the revision process could be extended. We agree with the most of the above remarks. Furthermore we believe that the revenues variable (for the corporates) without a high-granularity clustering (sector and geography) process could be uninformative. Hence we expect a substantial improvement of the quite fragile risk drivers set up for the corporates.

References

[1] Basel Committee on Banking Supervision (2014), “Revisions to the Standardised Approach for credit risk”, paper 307.

[2] Basel Committee on Banking Supervision (2014), “Fundamental review of the trading book: outstanding issues”, paper 305.

[3] Basel Committee on Banking Supervision (2014), “Guidance on accounting for expected credit losses”, paper 311.

[4] Basel Committee on Banking Supervision (2013), “The standardised approach for measuring counterparty credit risk exposures”, paper 279.

[5] Bonollo M., Marazzina D. (2014), “Lo Standardized Approach per Credit Counterparyt Risk”

[6] Bonollo M. (2014), “The Fundamental Review of Trading Book (FRTB) Revolution o (R)-Evolution? Innovazioni e Impatti”

[7] Gordy M.B. (2002), “A Risk-Factor Model Foundation for Ratings-Based Bank Capital Rules”.