Dutch Securitisation Association

Recital 27 provides as examples of homogenous pools “residential mortgages” or “corporate exposures”. We appreciate that EBA wants to refine this further (a combined pool of Dutch and Italian residential mortgages would not be regarded as homogenous by most investors, while individually these pools might qualify as homogeneous, subject to other criteria), but the approach taken in the consultation may lead to unnecessarily fragmented pools with insufficient diversification
We do agree with the criteria (a) to (d). As regards (d) we note that, apart from the relevance for certain asset categories, some criteria (jurisdiction, governing law) are far more relevant overall than others (type of obligor, type of credit facility) . Also, the determination which risk factors are relevant or not creates a lot of uncertainty and lack of clarity.
Apart from the comments on Q2, we fear that the fragmentation created by the risk factors will lead to smaller pools which will not be securitised because of lack of economics of scale. This defeats the purpose of the whole STS exercise.
We do agree that there should be a clear differentiation between ABCP and non-ABCP given the different structural support elements of both types of structures.
The most relevant risk factors should be equally applicable for all non-ABCP asset categories and pool specifics.
For Trade Receivable transactions, which form the core of ABCP programmes, (c), asset category classification, should be sufficient, since (d), risk factors, would only offer one, hardly relevant, risk factor, i.e. industry of the seller, for consideration.
We do agree with the list of asset categories. No further mergers are needed. Any missing categories (Project Finance, Whole Business Securitisation) are currently not relevant for the Dutch/European market and will not classify as STS.
Further specification of the definitions is not needed, as long as criteria (a) and (b) are also applied. It is relevant that assets are originated and serviced according to common standards, not that they comply with a theoretical definition.
We refer to our answers on Q2 and Q3.
We also note that in Art. 1 states that exposures “shall be deemed to be homogenous where they have similar risk profiles and cash flow characteristics”. If (g), type of repayment or amortisation, would be included as a risk factor, cash flow characteristics should be deleted from this general description (so leaving just “similar risk profiles”).
Including “similar cash flow characteristics” as a pre-condition for homogeneity implies that none of the asset classes as defined will be homogenous, since in all categories there will be a mix of amortising, revolving etc. exposures.
We do agree with the distribution of the risk factors generally. Although not all risk factors have the same weight, in our view, we appreciate that a further refinement or weighting of the factors would make the classification of homogeneity too complicated.
We generally do agree with the definition of governing law.
Prepayment changes over time, which cannot be captured by an up-front risk factor and prepayment differs per jurisdiction, which is (indirectly) covered by risk factor (i).
We do consider pools with predominantly second and lower ranking liens as a different risk category compared to pools with mainly first or equally ranking liens, so this risk factor may be relevant for determining homogeneity.
Yes, we do agree with this approach.
This depends on the relevant risk factor. If we would look at all risk factors proposed by EBA we would suggest indeed a materiality threshold for factors (a) to (g). For (h) Industrial sector of the seller, it is not relevant and for (i) and (j), Jurisdiction/Governing Law, it should not be difficult to select a 100% compliant pool.
IRB models should follow as much as possible the asset classes, and not the other way around, reason why we would like to limit the fragmentation of asset classes, since IRB models will not be available on a very fragmented scale.
We would prefer the alternative option, which would allow us, with adequate justification, to put aside certain risk factors.
Both alternatives are problematic to the extent that they require “similar cash flow characteristics” (see also our answer on Q8.).
Of the 13 public transactions in our jurisdiction (the Netherlands) in 2017, 3 might have suffered to meet the requirements anyway under both options. The reasons: either legacy portfolios, where selection options are limited, or transactions of small originators that have to combine different asset types in order to get to the required critical mass.
For the other transactions we assume that the both options might have worked except for the requirement of “similar cash flow characteristics”, which would have been problematic where in most cases amortising and non-amortising exposures were combined in one pool.
This hybrid option potentially offers a solution for our concerns around the different level of relevance/weighting of different risk factors. A limited number of very relevant risk factors (like “jurisdiction”) would be treated as “real” risk factors (as per your option 1) and the other (like “type of repayment/amortisation”) could be taken into account “in the underwriting”. This would provide, in our view, the optimal combination of limiting the risk factors at the pool level (and consequently improving the economies of scale) and still potentially taking into account the whole list of risk factors (and maximize the homogeneity).
Also we would expect this to be the best (or only) way to legitimately combine amortising and non-amortising exposures in one pool. But to avoidance misunderstandings: our preference for this option is not limited to its application to “type of repayment/amortisation”, but also refers to other risk factors.
The advantage would be less fragmentation of pools, so better economies of scale, while still safeguarding the use of proper risk factors. A potential disadvantage is that this hybrid option provides less certainty in interpretation (and increases the risk of disputes and penalties based on a perceived wrong interpretation).
The concept of taking the risk factors into account and reflected in the underwriting standards, methods and criteria may need some further clarification and elaboration in the final version of the RTS.
Rob Koning