• It should be specified that amendments to the underlying loans which are limited in magnitude and amount under the servicing agreement are not active portfolio management. These are changes typically affecting maturities (e.g. increasing the tenor by 1 year), the interest rate index and/or margin and sometimes providing payment holidays. These changes are allowed in most of cash deals, and are also known as “permitted variations”. They are granted by the servicer to the borrowers for commercial reasons or to facilitate the repayment of the debt.

• We do not agree that the sale of underlying exposures should be considered as active portfolio management, to the extent that it is done in accordance to the credit and collection policies of the bank. Especially in synthetic guarantees, the bank is often allowed the disposal of the assets, as long as it occurs in accordance to its credit and collection policies.

• It should be clarified that the eligibility criteria should not only apply at individual exposure level, but that there may also be eligibility criteria for the “replenishment portfolio” (i.e. all assets that are being replenished on a given date), especially in the context of synthetic transactions.

• EIF already provided a comprehensive response on the homogeneity criteria under a separate consultations

• It could be helpful to clarify that the following circumstances would still fall within the understanding of “periodic payment streams”
- Loans with grace periods
- Transaction documentation allowing for payment holidays

• NO LESS STRINGENT UNDERWRITING STANDARDS: We believe that reference to underwriting standard/policies would need to be complemented with a reference to servicing and collections policies. In fact, investors do not just focus on how the credit was granted, but also on how effectively and quickly it was collected in case of default. We therefore suggest expanding the concept to servicing and collections standards.

• MATERIAL CHANGES TO UNDERWRITING STANDARDS. We believe the following three concepts are key

1. the securitised exposures have been originated by applying underwriting standards that are not materially different compared to exposures that were not securitised. We therefore believe it is too much to specify a time window before the issuance of the securitisation over which the policies might have changes
2. (for revolving deals) the underwriting standards will not change in a way that is prejudicial to the interest of the investors ,or, if known, would have influenced the decision of investors as to invest in the securitisation.
3. (for all deals) the servicing agreement provides that the collection policies will not change in a way that is prejudicial to the interest of the investors, or, if known, would have influenced the decision of investors as to invest in the securitisation.

Finally, in our view, the proposed text does not include a workable concept of what makes a change “material”. In our view, a material change would occur if it, had it been known to the investor, would have influenced his decision as to invest in the transaction or not.

SIGNIFICANTLY HIGHER RISK: we believe that this criterion is not needed because the investor performs a due diligence on the transferred portfolio; moreover, it might be difficult to implement the check that EBA is suggesting (i.e. comparing whether the score of securitised exposure is worse than the average rating of non-securitised exposures).

CREDIT REGISTRY: We believe that there is a possible misinterpretation of Art. 20(11)b: While Art. 20(11)b refers to an exposure “at the time of origination” (i.e. when the exposure was granted to the obligor), text reference 48 of the draft guidelines refers to an exposure “at the time of the origination of the securitisation”. This might be interpreted as a requirement to re-check the obligor score with one or more credit registries again before including the exposure into the transaction. We understand that some originators only check the obligor status once, namely when granting the exposure.

On the basis of what we see in the market:
• We recommend increasing the allowed share of RV to 50%;
• as most of the investors already apply specific stresses/haircuts to cater for sale concentration over time, we recommend removing this concept;
• granularity could be more effectively addressed by making reference to the concept of Effective Number, which should be, in our view, equal or greater than 200

Most of the ABS transactions in the last years have been issued without any hedging, also because the interest rate mismatch in transactions where both assets and liabilities are indexed to floating indices is minimal. It should be clearly spelled out that a transaction can qualify for STS treatment because the available credit enhancement acts as mitigant.

We disagree that sequential repayment should apply within sub-classes. There is a number of transactions which feature a senior note divided into two classes of notes (for the sake of argument class A1 and class A2) that pre-enforcement rank pari-passu with respect to interest and sequentially with respect to principal, while post enforcement they become pari-passu with respect to principal as well. These are very common contractual terms and we see no reason why this kind of deals should fall out of the STS definition

We believe the interpretation here should be as flexible/wide as possible because this is really something for parties (originator/investors) to negotiate (we understand the regulator might want this in realtion to synthethics, but for cash deals it should really be up to parties)

We suggest clarifying the following in respect of the 4 points of the criterion:
(a) the deterioration can be measured with a simple gross/net default ratio
(c) as normally an originator would not measure the value of the underlying collateral, this criterion could be interpreted by making reference to a clean PDL
(d) this could be captured under a test aimed at measuring the cash on the accounts over the current balance of the portfolio

In our view it would make sense to allow for a shortcut for servicers who are also the original lender, because else there would be a redundancy with the requirements as to “Underwriting standards, originator’s experience (Article 20(10)).
The qualifications are focussed on the experience of the acting persons at the servicer, policies and procedures and risk management controls. Would it make sense to include “suitable operating systems” as well

To us paragraph 81 is not fully clear. It basically says that if the originator cannot provide the data, then other data (“provided by a rating agency or another market participant”) should be given to prospective investors. This however only, if “all other requirements of that article are met”, which means that (of course) data should be for “similar exposures to those being securitised”. Practically speaking, there may be significant differences in the performance in the data between originators and therefore a benchmark analysis has very little value. Should this then not better be ineligible for STS and paragraph 81 just simply deleted? -> If the originator does not have a sound track record of the assets, then it should not be STS?