While the political agreement reached at the end of May settled the primary questions the ABS industry had over the shape of future regulation in Europe, the compromise text approved in ECON last week still carried a few surprises.
One big surprise was the extension of the prohibition on self-certified residential loans from the STS criteria to securitisations in general, in Article 17, which covers underwriting standards. As worded, the ban would appear to rule out any new securitisations of self-certified loans: “Where the underlying exposures of securitisations are residential loans, the pool of those loans shall not include any loan that is marketed and underwritten on the premise that the loan applicant or, where applicable, intermediaries were made aware that the information provided by the loan applicant might not be verified by the lender.”
The provision was not featured in any previous draft, though the language is copied over from the STS criteria. Self-certs were widely prevalent in pre-crisis mortgage origination, and hence securitised pools, in the UK, though the practice is now banned for new lending. If left unchanged and interpreted strictly, it would prevent the (full) refinancing of most legacy UK RMBS, including many non-conforming RMBS that have been newly issued or refinanced since the crisis and were priced to call. According to Bank of America Merrill Lynch, GBP 14bn of such deals have call dates in 2019 (when the rule would take effect) alone.
While this provision may jeopardise some NPL securitisations, other provisions in the final text settle uncertainty for them that previous drafts had introduced. In particular the adverse selection protections have been modified and clarified, with a recital expressly stating that they “should however not prejudice in anyway the right of originators or sponsors to select assets to be transferred to the SSPE that ex ante have a higher than average credit risk profile compared to the average credit risk profile of comparable assets that remain on the balance sheet of the originator, as long as the higher credit risk profile of the assets transferred to the SSPE is clearly communicated to the investors or potential investors”.
The definitions of securitisation itself, and of both originator and sponsor, were tweaked from their current incarnations in the CRR. The compromise’s originator definition includes the “sole purpose” restriction pushed by the EBA to close a loophole in the risk retention rules, which did not appear in the Parliament’s version. While this was foreseeable, the sponsor definition was surprisingly loosened. Whereas the various drafts envisaged keeping the existing CRR definition of investment firms, which references MIFID and requires certain authorisations, the final version references MIFID directly, dropping the authorisation requirement. This should open up sponsorship to non-European fund managers, and firms which only have asset management permissions under MiFID, though not AIFMs. A&O warned in a client briefing however, that bringing non-EU entities in scope may not have been what the authorities intended to achieve.
The securitisation definition has also been changed, clarifying the grey area between specialised lending and securitisation, which had caused some ambiguity as to the classification of certain structures. Under the new rules, a securitisation “does not create exposures which possess all of the characteristics listed in article 147(8)” of the CRR, namely the specific features of specialised lending.
On the other hand, the text appears to impose new restrictions on who can be a servicer. A recital specifies: “A sponsor should be able to delegate tasks to a servicer, but should remain responsible of the risk management. In particular a sponsor should not transfer the risk retention requirement to his servicer. The servicer should be a regulated asset manager such as a UCITS manager, an AIFM, or a MIFID entity.”
It is not clear, however, where in the operative text this requirement is imposed. The definition of servicer references CRR, but that definition only states that “’servicer’ means an entity that manages a pool of purchased receivables or the underlying credit exposures on a day-to-day basis”. Instead the sponsor definition talks about delegating day-to-day portfolio management to an authorised entity.
The STS criteria require the servicer to “have expertise in servicing exposures of a similar nature to those securitised and shall have well documented and adequate policies, procedures and risk management controls relating to the servicing of exposures”, but do not limit the role to regulated entities.
One senior securitisation lawyer argued this restriction is likely only intended to restrict sponsors, particularly in the context of ABCP rather than being a general requirement for securitisation servicers. On the other hand, A&O’s commentary points out that under the existing definition, sponsors have been deemed suitably responsible regardless of whether or not the delegated servicer is EU authorised.
The final text also adds a definition for “original lender”, which was not previously spelled out. It is based on the definition of originator and covers the entity which “itself or through related entities, directly or indirectly, concluded the original agreement which created the obligations or potential obligations of the debtor or potential debtor giving rise to the exposures being securitised”.
The text sheds light on some of the compromises reached between the co-legislators. For instance, there remains a provision to compare losses in securitised portfolios against those remaining on originators’ balance sheets, but it is greatly watered down from the Parliament’s proposal. Instead of requiring originators to disclose one-year loss rates, the final texts enables competent authorities to impose sanctions if they find evidence that losses over four years, or the life of the transaction where shorter, are “significantly” greater on securitised pools than on balance sheet because of the “intent” of the originator.
Similarly, the final text’s sanctions regime moves away from the Commission’s proposed strict liability and calls for punishment only when breaches are due to “negligence or intentional infringement”. Another loosening of the oversight regime grants a three-month grace period if an STS notification is discovered to have been made erroneously but in good faith. During this period the originator, sponsor and SSPE may rectify the problem and the deal will continue to be considered STS. On the other hand, the possibility of criminal sanctions with large fines in case of deliberate breach, and even prosecution of individuals, remains.
Another visible compromise concerns the restrictions on choice of SPV jurisdiction. The final text does not include a third country equivalence regime for STS securitisations, as Parliament wanted, but for all securitisations, SSPEs may not be established in third countries which are listed as a High-Risk and Non-Cooperative Country and Territory by the Financial Action Task Force on Anti-Money Laundering and Terrorist Financing, or where they have not signed an agreement with a member state on tax governance and exchange of information.
CMBS explicitly banned
Despite minimal discrepancies between the various drafts’ STS criteria, other than for ABCP, the final text does contain several material changes and a few surprises. The explicit ban on CMBS envisioned by the Council is included, though the allowance of “business property” as a collateral type, as in the Parliament text, may provide a loophole. Residual values backed by a repurchase obligation are explicitly allowed, addressing the auto industry’s fears over the prohibition on repayment depending “predominantly” on the sale of assets. Unlisted corporate bonds are eligible — possibly a concession to SME securitisations backed by Schuldscheine.
The disclosure requirements show signs of both loosening and tightening. The requirement to disclose all essential underlying documents is tempered by a caveat that “a summary of the concerned documentation” can be provided to avoid breaches of privacy law or confidentiality obligations. On the other hand, originators and sponsors must provide a liability cashflow model “which precisely represents the contractual relationship between the underlying exposures and the payments flowing between the originator, sponsor, investors, other third parties and the SSPE, and shall, after pricing, make that model available to investors on an ongoing basis and to potential investors upon request”.
Private transactions are exempted from the requirement to notify their disclosures to a securitisation repository, but are otherwise still required to disclose the same information to investors as public deals.
Similarly on the due diligence side, the requirements largely reflect the Commission’s streamlined list, removing the assessments required under some existing regulations of originators’ and sponsors’ statements about due diligence and collateral quality, and of the collateral valuation methodology. In a variation from the Commission proposal, for fully supported ABCP programmes, investors in the CP can focus their due diligence on the structural features of the programme and the strength of the liquidity support, rather than the underlying assets.
Some of the details emerging from the text were trailed previously after the political agreement, but only at high level. For instance, the carve-outs from the ban on resecuritisations are elaborated. Deals issued before the date of application are exempted, as are those used for specified “legitimate purposes” as approved by an entity’s competent authority, which until further specified by ESMA are “the facilitation of the winding up of a credit institution, an investment firm or a financial institution, ensuring their viability as going concern in order to avoid their winding up” and “where the underlying exposures are non-performing, the preservation of the interests of investors”.
In addition, fully supported ABCP programmes are not considered resecuritisations “provided that none of the ABCP transactions within that programme is a resecuritisation and that the credit enhancement does not establish a second layer of tranching at the programme level”.
In the CRR text, the full details of the new hierarchy of approaches is apparent. Generally, the Standardised Approach (SA) will rank ahead of the Enhanced Ratings Based Approach (ERBA), resulting in significantly lower capital charges than under the Basel hierarchy for senior tranches in peripheral jurisdictions, for example.
Mezzanine tranches of UK prime RMBS would also be a major beneficiary, according to BAML research, in some cases seeing a drop from existing capital charges.
There are various “protections”, however, to limit the potential for arbitrage, and auto loans and leases are subject to a specific carve-out (under the standardised approach, the underlying exposures attract a 75% risk weight, far higher than for residential mortgages).
Use of the SA would also be disallowed for STS securitisations where it would result in a risk weight higher than 25%, or for non-STS securitisations where the SA would result in a risk weight higher than 25% or the ERBA would result in a risk weight higher than 75%.
In addition, the final CRR text broadly adopts the calibration proposed by the Basel Committee in its latest revision of the securitisation framework, rather than those previously put forward by the co-legislators, including the same reduced floors and ERBA look-up results for STS securitisations (STC in Basel’s jargon). This results in higher capital charges across the board compared to the existing calibration, but STS securitisations, particularly at the senior level, have lower floors and absolute risk weights than their non-STS counterparts.
The industry may breathe a sigh of relief after the proposed amendment to EMIR to grant STS SSPE’s equivalent treatment to covered bond issuers — which would exempt them from clearing and posting two-way margin — made it through the trialogue process after being thrown in doubt by theCommission’s proposal to reclassify securitisation issuers as financial counterparties. What the future holds for non-STS issuers remains unclear, however.
Despite the extra clarity provided by the legislative texts, much remains to be specified and clarified. The reforms call for a slew of RTS to be produced, mainly by the EBA and ESMA, within six or 12 months of the date of entry into force of the regulations, which should in principle allow them to be adopted in time for the delayed date of application, 1 January 2019. In practice, however, the workload will be very challenging. The subject matter includes risk retention, transparency obligations, due diligence requirements, homogeneity of assets in STS pools, ABCP criteria, third party certification, EMIR requirements, use of proxy data in the Internal Ratings Based Approach, and three separate RTS on data repositories.
Key areas where the intent and reach of the final text remains unclear include the status of otherwise eligible third country entities under risk retention rules, whether the carve-outs for resecuritisations apply to capital charges under the CRR as well, the consequences for investors of breach of the due diligence requirements. The Commission must also publish its delegated act on Solvency II, outlining the capital treatment for STS (and non-STS) securitisation investments by insurers.
The compromise text is not technically final, as it is subject to linguistic and legal review before the plenary vote scheduled for October. Lawyers spoken to by Debtwire do not expect major changes, however, given the limited remit of those reviews.