By Robert Paterson, ABS Research, Morgan Stanley Dean Witter
The Basel Committee published a second consultative package on the New Basel Capital Accord on Jan 16. Readers will remember the committee is working to modernize the current system of risk weightings that is contained in the 1988 Basel Capital Accord. The package develops the theme of a two stranded approach to calculating capital requirements: either according to a set of standardized ratings based risk weightings, or the bank's assessment of the exposure's default risk, as was outlined in the first set of proposals entitled A New Capital Adequacy Framework, June 1999.
Perhaps more significantly for the securitization market, the New Basel Capital Accord also provides extensive, although not complete, guidance on the treatment of capital relief for banks securitizing their assets. The most controversial of these proposals is the potential ex ante capital charge on all securitizations that may be levied against issuers to cover residual and implicit risks.
The Two Stranded Approach
The second package retains the two methodologies for the assessment of capital charges, i.e., the standardized approach and the internal ratings based (IRB) approach that were outlined in the 1999 paper. However, the second package goes into detail as to how internal credit models would be expected to work. It also contains a wider definition of the banks it expects to become users of the IRB approach having moved from sophisticated banks to those that are either internationally active, or have above average risk profiles.
The standardized approach involves the use of rating agency ratings to determine the risk weight to be applied to a banking exposure.
The IRB approach is further split into a foundation and advanced approach. Those who follow the foundation approach will be allowed to subject to approval of their internal models use their own assessments of probability of default for assets, although they will be required to continue to use their regulator's estimates of loss given default. Generally, those who follow the advanced approach will be able to use both their own estimates of the probability of default as well as the loss given default. However, this is not the case for ABS investors who may be required to use a 100% loss severity assumption regardless of whether they are using the foundation or advanced approach. This would mean there would be no difference between the foundation and advanced approach for ABS investors.
From the Investors'
Perspective
From the perspective of investors in securitized products, the proposals under the standardized approach are unchanged from the original 1999 proposals. If implemented, banks with holdings of highly rated ABS that adopt the standardized approach will see a significant reduction in the amount of capital they are required to hold against these positions. This should all else being equal have a positive effect on the spreads of highly rated ABS paper.
Users of the standardized approach that purchase the senior tranches of unrated deals may be able to use a "look-through" approach. This allows them to potentially achieve a risk weighting for the bonds that is equal to the highest risk weighting of the assets within the securitized pool. Therefore if one had a portfolio of 20% risk weighted assets, the weighting would be 20%, although if there were a single asset with a 100% weighting within the pool, the senior tranche of an unrated pool would achieve a 100% risk weighting based upon the "look-through" approach. Subordinated and mezzanine tranches of unrated bonds can potentially be 100% risk weighted, if the investing bank's regulator agrees the tranche would be likely to receive an investment grade rating if it were rated. If, however, the regulator believes the tranche would receive a non-investment grade rating, the tranche would attract a one for one deduction from capital because it would be deemed to be a form of credit enhancement. Investors who operate under the IRB will either have to use the probability of default of the rating agency for the given rating, or use their own model's probability of default for the bank's internal rating that is equivalent to the external rating. More significantly, although banks would be allowed to use their own assessment of probability of default under the IRB, it is proposed that they will not be allowed to use their assessment of loss given default for securitized products under the advanced IRB treatment. In effect, this will make the foundation and advanced approaches the same for securitized products. Even worse is that the Committee has proposed that banks should use a loss severity assumption of 100% for securitized products compared to a 50% assumption for sovereign and corporate debt. Whilst the committee note that this appears conservative, we think this treatment is especially harsh. We have calculated that a bank would be better off using the standardized approach when assessing the capital charge at both highly rated and lower rated pieces of ABS (see chart). Given that the stated intention of the Committee is to provide an incentive of a reduced capital charge to users of the IRB approach, this may be changed before the final guidelines are published. Indeed, we note this is one of the specific areas on which the Committee have asked for industry comment.
For users of IRB who invest in unrated bonds it appears as if the investing bank would be expected to make a deduction from capital. This is because it is assumed unrated tranches are in fact forms of credit enhancement. However, the committee suggests that this issue should be examined further, and could potentially be tackled by allowing either an implied rating of the unrated tranche to be used, or alternatively allowing the investor to determine, through due diligence, how much capital they should hold if the entire portfolio were on their balance sheet and then from this, calculate the capital charge to be applied against the tranche.
From the Issuers' Perspective
Under the new guidelines, for banks to receive an off-balance-sheet treatment for capital purposes of securitized assets, a "clean break" must be achieved. To achieve this the bank must have transferred the assets via a true sale - synthetics are discussed later - which means:
*the assets are bankruptcy remote from the seller;
*the transferee must be a qualifying SPV and the holders of interests in the entity must have the right to pledge or exchange those interests;
*the transferor must not maintain control over the transferred assets - servicing assets is not constituted to mean having control of the assets.
In addition clean up calls should be limited to a relatively small proportion of the overall issuance. This level is not defined in the paper, however if the issuer wishes to maintain a call above "a relatively small percentage" or call the deal above the clean up call, they can only do so with permission of their national regulator.
Under the standardized approach issuing banks will have to deduct credit enhancements they provide from their capital bases. If the first loss piece is of sufficient size, issuers may be able to secure a risk weighting for second loss tranches based on its external debt rating. Revolving securitizations with early amortization triggers will carry a capital charge of between 10%-20% of the investors' interest. The percentage will be set at the national regulator's discretion and will depend on a number of factors including amortization provisions and the terms of any clean up calls. This is a reduction from the suggested 20% of the original capital charge of the entire portfolio that appeared in A New Capital Adequacy Framework. Short-term liquidity drawings will be risk weighted at 100%, although undrawn facilities will be given a 20% risk weighting. If sponsor banks providing liquidity facilities to ABCP (asset backed commercial paper) conduits can show their facilities are genuinely to provide liquidity, they can assign the same rating. This will markedly increase the cost of these lines to the conduits and we expect this to curtail their business unless revisions are made to this guideline. The committee is especially keen to punish institutions that are found to have provided "implicit recourse" to transactions (i.e., supported the transaction beyond their contractual obligations). Consequences mentioned in the paper are the potential loss of capital treatment for assets within the structure or for all securitized assets, as well as the removal of the ability to achieve off balance sheet treatment for any future deals. In addition, the bank may be forced to disclose to the markets the negative affect on their capital position, and their inability to securitize assets in the future. The committee also mentioned in a single paragraph the possibility of an ex ante capital charge to take account of the residual and implicit risks in securitization. The charge would be assessed on a deal by deal basis according to the national regulator's perception of the residual and implicit risks to the issuer within the transaction. Without more detail, it is difficult to comment on this issue, although we are sure the industry will fight any application of such a charge.
Synthetics
The committee specifically mentioned synthetic structures as an area for further work. They feel it appropriate that some form of charge be made for the risk that the credit default swap arrangements in these transactions may be mismanaged or that another form of operational or reputation risk may affect the originating bank. The committee also noted that institutions using the IRB approach are less likely to use synthetic structures going forward, as under the new guidelines regulatory capital will be brought closer to economic capital usage, thereby reducing the incentive to enter into securitizations simply to reduce capital requirements.
Implementation
In addition to requesting market comment on specific points of interest, the Basel Committee has invited comments on the entire set of proposals by May 31, 2001. The committee then envisage producing final guidelines by the end of 2001, with implementation at the national level to follow in 2004. The committee aims for the guidelines to be, in so far as is possible, implemented at the same time across jurisdictions in order to reduce the possibilities for regulatory arbitrage. Readers should note that the US regulators have agreed to postpone the implementation of the new accord until 2004. The delay between the end of 2001 and 2004 is to allow national regulators time to pass the guidelines through any legislative processes that are required and to prepare to be able to implement them. Within Europe the guidelines will first have to be incorporated within an EU Directive, and then be implemented by individual national regulators. Whilst the guidelines are not binding on national regulators, we would expect them to be broadly applied as many are members of the Basel Committee and all have been given the opportunity to submit their comments and shape the proposals.
One area where we have greater skepticism is in the ability of the regulators in some jurisdictions to resource the assessment of internal capital models adequately. Whilst the Committee has mentioned its intentions to encourage dialogue between different regulators, we are not convinced that there will be an equal implementation of the rules across jurisdictions.
Conclusions
We believe there is a substantial amount of work that needs to be done in relation to securitization before the Basel Committee should publish final guidelines. In particular, we think the following areas require considerable attention, and/or will be lobbied against by banks who both issue and invest in securitized products:
*an ex ante charge for residual and implicit risks within Securitisation for issuers;
*the proposed capital charge for issues of revolving assets which have early amortisation provisions;
*the imposition of a 20% risk weighting on one year and under liquidity facilities. This is especially detrimental to ABCP conduits;
*under the IRB approach the committee admits that a loss given default assumption of 100% is conservative. We would agree this is especially so in that a 50% loss severity has been proposed for both sovereign and corporate debt.
On a more positive note we welcome the often mentioned reduced risk weightings that are proposed under the standardized approach for ABS, as they should provide a positive influence for the spreads of highly rated ABS paper.
