The aim of the regulations (STS and CRR amendments) is primarily to diversify the financing sources of European enterprises and to place the allocation of risks in the European financial system on a broader basis. Besides, removing the burden from banks’ balance sheets through the intended intensification of securitisation activity is to enable a further expansion of lending to the real economy. We consider this context important in analysing the present draft regulations.
In our view, the Executive Summary of the draft RTS correctly sets out: ”The overarching objective of the homogeneity requirement is, ……, to simplify and facilitate the assessment of underlying risks by investors and to enable investors to perform robust due diligence” (S. 4, item 2). We also maintain that the overarching objective of the homogeneity requirement should be directly related to investors due diligence.
It should be taken into account, however, that ABS investments are always made by professional investors – this also applies explicitly to investments in STS securitisations – and that the sale to inexperienced retail investors is ruled out under Article 3 of the Securitisation Regulation.
Professional ABS investors have effective instruments and processes in place to analyse and assess securitisations. Consequently, they have to rely on high data transparency and access to information on the relevant risk factors, underwriting and servicing standards in order to be able to perform robust and well-founded due diligence. In the course of the financial crisis, lessons were already learned for improving regulation and the transparency requirements in connection with securitisation transactions were tightened considerably. In the context of the current Securitisation Regulation, the existing transparency requirements have been developed further once again.
Compared with transparency, the degree of ‘homogeneity’ of the underlying portfolio of receivables plays a less important role for professional investors in terms of analysis effort and expressiveness of results. However, excessively tight and restrictive homogeneity requirements may lead to undesirable side-effects with regard to volume, risk profile and transaction costs of STS-compliant portfolios of receivables. We will address these side-effects in greater detail below as part of our answers to questions Q1 to Q3.
The current Securitisation Regulation and the homogeneity requirements contained therein for STS securitisations therefore need to be considered in close connection with the transparency requirements. After all, the ability of an investor to perform robust due diligence increases precisely as a result of the investor’s good compliance with the transparency criteria.
Q1: Do you agree with the focus of the RTS, general approach and underlying assumptions on which the RTS are based? Does the proposed approach provide sufficient clarity and certainty on the interpretation and application of the criterion of homogeneity?
In our opinion, the approach of the draft RTS can provide a basically suitable starting point for the formulation of reasonable homogeneity requirements. However, particularly the risk factor approach set out under Art. 1 (d) in conjunction with Art. 3 of the draft RTS can lead to a material restriction of potentially homogeneous portfolios. This approach translates the general requirement of the Securitisation Regulation (similar risk profiles and cash flow characteristics) into numerous possible individual requirements, but these can hardly be fulfilled by specific portfolios in their entirety.
The question to what extent homogeneous portfolios can be put together at all in future thus depends crucially on how portfolios with a certain residual heterogeneity can be assessed. The note in Recital 12 of the draft RTS, according to which the institutions select the relevant risk factors from the factors to be taken into account that are mentioned in the draft RTS and assess them for their impact on “similar risk profiles and cash flow characteristics” should, in our view, be detailed and explained in greater depth. Considering the drastic sanctions provided for under Art. 32 of the Securitisation Regulation, we consider it urgently necessary to provide more clarity on the implementation of the homogeneity requirements, particularly with regard to the risk factor approach.
We consider the requirements introduced into Article 1 (a) and (b) of the draft RTS for the underwriting and the servicing to be acceptable provided they are not applied too restrictively.
The reference to the asset categories (Article 1 (c) of the draft RTS) was already included in the Securitisation Regulation.
The Securitisation Regulation does not demand risk factors to be applied (Article 1 (d) of the draft RTS) to operationalise the objective of achieving similar risk profiles and cash flow characteristics, but we can comprehend the motivation of the EBA to pursue this approach. Nevertheless, we see the danger here that a narrow interpretation by the regulator may lead to the risk factor approach being applied too restrictively in practice and hampering the fulfilment of the objectives of the Securitisation Regulation. This applies in particular to securitisations of the asset categories “corporates and SMEs”, but also to “auto ABS” and “RMBS”. Further remarks on this under Q3.
In our view, the homogeneity criteria mentioned harbour the danger of too narrow regulatory interpretation so that the number and scope of homogeneous portfolios of receivables would be materially restricted:
1. Recital 43 of the Securitisation Regulation explicitly emphasises that the design of the RTS with respect to the issue of homogeneity must ensure “that the securitisation of SME loans is not negatively affected”. We are of the opinion that the extensive list of risk factors to be taken into account for the asset category “Corporates and SMEs” potentially runs counter to this objective of the EU Commission, the Parliament and the Council. The structures of the requirements of this asset category reflect the wide range of company-specific financing structures of corporates and SMEs in Europe. This spectrum is broader than the narrow understanding of “homogeneity” of the draft RTS. We expect that, for corporates and SMEs, the excessive narrowing of homogeneity requirements down to “similar risk profiles and cash flow characteristics” leads to portfolios of receivables that are consistently homogeneous and marketable from an economic perspective only in exceptional cases.
2. We are concerned that – contrary to what has been set out in Recital 3 of the draft RTS – the risk factor approach will not always lead to adequately diversified portfolios. The portfolios that are homogeneous in the understanding of the draft RTS may turn out to be granular in terms of borrower-specific risks, but they will always have a high concentration of risk with regard to the relevant risk factors because of the way they are designed. Risk concentrations fundamentally increase the performance volatility of ABS bonds and, hence, the direct and indirect costs of the transactions.
3. The targeted composition of securitisation portfolios in accordance with the criteria of the draft RTS may create a situation in which the risk profiles of the securitised portfolios and those of the receivables remaining on the balance sheets of the originators differ systematically. Thus, it cannot be ruled out that the homogeneity requirements of the draft RTS will systematically run counter to the requirements of Article 6 para 2 of the Securitisation Regulation.
The importance of homogeneity in analysing the portfolio of receivables therefore is subject to a cliff effect: first, as homogeneity increases, the advantages initially predominate in the analysis and understanding of the risk profile. From a certain degree of homogeneity onward, however, the disadvantages mentioned predominate. We therefore suggest that the homogeneity requirements should always be analysed with a measured view and always taking into account the data and document transparency of each specific case, namely with an explicit view to the possibility of performing the required robust investor due diligence with methods and processes customary in the market. This could be done along the lines of a comply-or-explain procedure:
• comply: To the extent that meeting the requirements specified in the draft RTS enables sufficiently homogeneous overall portfolios that are marketable from an economic perspective, the institutions would apply the provisions of the draft RTS for the entire portfolio of receivables.
• explain: To the extent that a marketable portfolio can be put together only by accepting a certain residual heterogeneity, this should be innocuous in the context of the homogeneity requirements provided that
o all receivables belong to the same asset category,
o the portfolio is homogeneous with respect to the requirements of Article 1 (a) and (b) of the draft RTS,
o the portfolio contains residual heterogeneity only with respect to one or few relevant risk factors,
o sufficiently transparent data and documentation are available to analyse the portfolio.
In their analysis and assessment of ABS investments, professional investors mainly focus on the past performance of the relevant asset class. So the main requirement for a robust due diligence process with customary procedures is that all the data and information necessary for analysing a portfolio of receivables be made available. With respect to the homogeneity of the receivables, it is sufficient that the portfolio be largely homogeneous to enable a smooth analysis and assessment.
Having said that, we would like to explicitly point out that this proposal does not mean exchanging material “un-homogeneity” for “transparency”, but aims to make the natural residual heterogeneity of a portfolio of receivables accessible for the required robust investor due diligence using methods and parameters that are customary in the market.
We also believe that the significance of a risk factor must be viewed against the background of whether it is predictable or not for the investor on the basis of historic data.
In static and highly granular profiles, for all risk factors it is relatively easy to calculate what influence they have on the development of the portfolio, and where historic data are available that can be applied to the portfolio and clearly illustrated in cash flow models and investor reports. Hence, to the extent that the possible risk factors are known and sufficiently easily predictable on the basis of historic data (“known known”), they should be regarded as not very relevant for the purposes of determining the homogeneity criteria.
However, there are also risks that are not predictable, neither with respect to their type nor with respect to the intensity of their occurrence. The collapse of the Soviet bloc and the attacks of 9/11 belong to this category of risks. These unknown risks (“unknown unknowns”) cannot be taken into consideration as relevant risk factors either.
Influencing factors also exist, on the other hand, that are known today in terms of their nature but whose future development is unpredictable.
This could be the case especially in revolving and/or non-granular portfolios. (“known unknown”). In our view, such risk factors should be regarded as relevant and therefore be taken into consideration additionally as a homogeneity criterion.
The application of the relevance definition set out above can be explained using the following example:
Let us assume a leasing transaction in which the leasing payments and guaranteed residual values are mixed. While the lessees belong to a specific type of obligor (for example SMEs or individuals), the residual values are based on the manufacturer’s guarantee (such as a large corporate). In addition, the leasing payments and residual values have different amortisation patterns. The risk factors “Type of Credit facility”, “Obligor Type” and “Type of Repayment” must therefore be assessed for their relevance. If the replenishment conditions fail to set out the range in which the proportion of residual value in the portfolio can fluctuate over time, the ratio between leasing payments and residual value is unpredictable (“known unknown”). Hence we would be dealing with several relevant risk factors that could materialise in such a portfolio and would have to be considered in assessing homogeneity.
If, on the other hand, the above example includes a contractually agreed and sufficiently narrow corridor within which the proportion of residual value may fluctuate, it would again be possible to make a forecast of cash flows with a view to contractual, default- and prepayment-related characteristics (“known known”). Consequently, in this case the specified risk factors would be considered not relevant for assessing homogeneity.
With regard to the question whether a portfolio is “homogeneous” in the meaning of the draft RTS, in other words, whether the assets have similar risk profiles and cash flow characteristics, the focus is on the characteristics of the assets and not on the type of transaction. The type of transaction should therefore not be given any particular importance with regard to homogeneity requirements.
The list of asset categories is sufficient in our view. However, it should be clear that residential mortgages may include not just the owner-occupied house or apartment, but may definitely include tenanted objects as well, provided the landlord is a private individual who rents out their premises on a non-commercial basis.
We do not believe that it would make sense to further specify SMEs in accordance with the Commission Recommendation 2003/361/EC because this Commission Recommendation primarily fulfils statistical purposes but is not consistent with the practical design of the European business and securitisation markets. An additional specification of the SME segment by the Commission Recommendation would further restrict the scope for homogeneous portfolios of SME receivables.
We generally agree with the approach of determining the actually “relevant” risk factors for a specific portfolio from a defined list of risk factors “to be considered”. In this regard we presuppose that the institutions independently determine and document their procedure for selecting and assessing the relevant factors.
We regard the risk factors to be considered as sufficient.
We agree, a further specification of the governing law does not make sense – at least not from a German perspective.
No. In Germany, prepayments are also tied to a legally regulated prepayment fee which is included accordingly in the cash flow and covers the financial loss for the investor in full (residential mortgages) or in part (consumer loans). Consequently, they do not play an important role.
Both questions can be answered with Yes with regard to the asset category of residential mortgages; however, this does not play a role for the asset categories of auto and consumer loans. In the case of SME and consumer loans the collateral is usually not assigned to any particular loan but is used as security for all loans of a borrower.
Yes, a uniform materiality threshold should be introduced for each risk factor. In our view, this threshold should be 10%. We view the application of a threshold as extremely important particularly for the asset category “SME and corporates” pursuant to Article 2 (d) of the draft RTS.
With respect to homogeneity, we see no merit in using elements/factors of IRB modelling. This would mean linking two largely independent regulatory thematic areas. In our opinion, it would generally increase the complexity of implementing regulatory requirements without leveraging any synergies.
We regard Option 1 as the better option as it provides more clarity in regard to the risk factors. However, we believe the following two points need more clarification.
First, it should be explained more clearly how the term ‘risk factor’ is to be understood (see our answer to Question 4) and whether it also includes the risk factors that are self-explanatory in terms of their impact on the portfolio through historic data.
Second, if the answer to First is Yes, it should by all means be permissible to use the comply-or-explain procedure which we proposed under Question 3.
No, this would make it unnecessarily complicated and unclear.
See our answer to Q18 above.
See our answer to Q18 above.
We are convinced that our approach is logical and conclusive. We are not familiar with all the answers from other institutions, but of course we know the answers from the automotive industry that we would support if EBA did not share our solutions.