German Insurance Association (GDV)

We would like to add to that list of impediments the importance of the investor's internal capabilities in evaluating, measuring, and managing ABS risk. Before the financial crisis, many market participants had internal risk-management measures and models in place, although reliance on external ratings was certainly strong. We think that credit work has been further intensified since then - on the one hand self-imposed to better understand risks and on the other hand driven by increased regulatory requirements on the supervision of asset-backed securities. The extra work and costs involved may, however, to some extent be another reason why some investors/banks are reducing their activity in securitization markets.

Securitization as such is an asset class that perfectly meets the needs of insurance companies with regard to yield enhancement, diversification, measurable risk, structural enhancement and sustainable cash flows. Therefore, securitization should not be punished with inappropriate capital charges compared to other credit asset classes.
We agree with the recommendation to exclude synthetic securitization from the framework for simple standard and transparent securitization. But we think it should be carefully analyzed to exclude them completely, because synthetic transactions could become relevant for risk hedging.
Yes, we find the default definition under Criterion 5 (ii) appropriate. We highly welcome a standardized default definition in general. We deem a period of 90 days past due as appropriate to account for defaulted assets. This definition should apply to all kinds of asset types with no exceptions (securitizations from southern Europe tend do have longer terms until being defined as defaulted). No difference should be made, however, between the definition of defaulted assets within the Standard Approach and the IRB Approach. Additionally, we would require no delinquent underlying exposures at the time of inclusion in the securitization.
We believe there should be no restrictions to the type of jurisdiction. The investor should be free to choose which country he wishes to be exposed to risk in. However, it should be clear where the assets originated, the country of governing law/regulation, which jurisdiction the originator is based in and where the SPV was established.
The senior class usually makes up the largest part of a securitization, but we think that a distribution of all voting rights towards the most senior noteholders might deter investors of non-senior tranches. The motivation for voting is definitely different or even higher for mezzanine/junior investors, which are exposed to higher risk. We believe that the specific interests of different investors should be respected. In our opinion a simple distribution of voting rights to protect the credit quality of the senior class would have an adverse effect. As long as the senior class is well covered there is no motivation for the senior class to take action whereas junior and lower mezzanine classes have an intermediate interest to keep in the portfolio losses low. If however a certain over collateralization for the senior class is not guaranteed anymore the senior class should receive the rights to take action.
All relevant underlying transaction documentation should ideally be disclosed prior to pricing/issuance. Due to the fact that analyzing securitization involves extensive credit work (from an investor's perspective), we would welcome a period of minimum five days between disclosure of all relevant documentation including loan-by-loan data tape and pricing/issuance.
We believe that granularity is one of the most relevant factors for determining the credit risk of the underlying. However lower granularity does not lead necessarily to higher losses. Credit risk should not be measured by granularity alone. Looking at the US RMBS market, most subprime RMBS pools were granular per se, but in the end showed severe default rates. The regulator should also bear in mind that other factors are equally important when measuring credit risk, for instance structural enhancement. Pools that tend to be less granular usually have higher protection in terms of structural enhancement to account for higher credit risk.

The granularity of the underlying portfolio depends strongly on the assets class. Hence, a general threshold is in our opinion a too imprecise limit and redundant. We, think that a professional investor is able to determine the credit risk and evaluate factors like granularity by itself. In this context we welcome the disclosure of loan tapes that include all important information regarding the underlying credit risk. It should be up to the investor to evaluate different credit risk factors like granularity, etc.
Criterion 2

We do not agree with the general exclusion of securitisation that allows active portfolio management because CLO transactions would not comply and be punished via higher capital ratios. There should be a distinction between CLOs and CDOs or CBOs. Most CLOs meet a real financing need of high yield corporates whereas CDOs/CBOs have been created for (rating) arbitrage purposes. Default rates in European CLOs have been very low over the past years compared to structural credit protection (the highest S&P ELLI default rate was 8% versus typical most senior credit enhancement of 30% and more). There are plenty of transactions that did not have a trading bucket in which default rates were clearly higher (for instance US subprime pools, Spanish RMBS). With regard to CLOs, the collateral manager is now obliged to retain a minimum level of 5% of the deal in order to show alignment of interests. Additionally, the trading bucket has to comply with certain eligibility criteria in order to guarantee a certain level of quality.

We would also like to highlight that the corporate issuers within CLOs are transparent, whereas in SME transactions, the pools are more granular, but the corporate is generally unknown. Leveraged loan CLOs include far less obligations (~80-120) than SME CLOs (several thousand). The exposure of SME CLOs are small cap companies (revenue < EUR 1m) for which information is hardly available. On the opposite leveraged loan CLOs have a line by line reporting of the loans in the portfolio on a monthly basis. The majority of companies are large enough to be rated and data/news are available. The companies are managed by PE sponsors and have even monthly accounting closings. All in all, we would like to have actively managed securitisations (CLOs) to be included in the simple standard and transparent securitisations framework assumed they fulfil certain disclosure minimum standards like easily accessible investor reports and loan-data tapes.

(please see graph 1 of the attached file)

Criterion 4

Under this requirement, some commercial mortgage-backed securities, master trusts, securitisations with embedded call options (clean-up call, optional call) and securitisations of residual values would not comply with the simple standard and transparent securitisations framework. As an investor we would, however, like these asset types to qualify and thus have preferred regulatory treatment in general. Clean-up calls make sense, since transaction costs become too high once the deal has amortized down below the usual 10%.

Moreover, we see the following points of improvements with regard to specific criteria for obligations:
• The example of CLOs shows that for underlying exposures which depend (partially) on the re-financing of the obligor the performance is better than comparable corporate bonds exposure. Further we criticise that a stronger structural enhancement of a securitisation with lower asset quality exposure has been disregarded with this requirement;
• In our view the refinancing risk can be limited if maturities are staggered in the portfolio.

Criterion 7

We, as an investor, would very much welcome a standardised disclosure of the CRR retention rule (where type of retainer, form of retention, level of net economic interest retained, and statement regarding on-going retention are clearly disclosed).

We would appreciate a reference on Bloomberg for all risk retention compliant transactions.

Criterion 10

Please further define the difference between revolving period" and "reinvestment period". We assume that revolving period includes the term reinvestment period.

Criterion 13

We do not believe this concept is necessary for all transactions and it does not necessarily lead to a simplification. Also, it is hardly possible to treat all investors of a transaction equally, since senior noteholders have different interests/motivations than junior noteholders. We think the ordinary stakeholders should be able to work out any conflicts of interest.

We prefer a clear definition of investor rights in the documentation. The inclusion of an external ‘identified person’ could lead to delays in restructuring processes.

Criterion 20

“Substantial similar” exposure is for us an undefined term. We would support that exposure which is comparable to a regional index is “substantial similar” (i.e. S&P European Leverage Loan Index).

Comparison SII:
• In comparison we are not in favour of the approach in the delegated regulation (Solvency II) only credit facilities to small and medium sized corporates are eligible for type I securitisations (Art. 177 No. 2 h iii). This would exclude loans to larger corporates.
• Further believe that the liquidity requirements in Art. 177 No. 2 b ii are not necessary to establish simple and transparent securitisations. In addition the liquidity is hard to measure (e.g. ES RMBS or SME CLOs).
• We appreciate that the EBA definition for high quality securitisation extends over the complete capital structure. This approach is more favourable than the Solvency approach where only the most senior tranche can be classified as type I securitisation (Art. 177 No. 2 c). This would mean that even AA rated second pay tranches are treated as type II assets if the senior tranche has not fully paid off.

In general, we do agree with the proposed credit risk criteria, but would like to comment on the following:

Criterion B:

We think that regulation here by including a granularity threshold is redundant. We very much welcome the disclosure of loan tapes and historical data that include material information regarding the underlying credit risk. It should be up to the investor to evaluate different credit risk factors like granularity, etc.

Criterion C:

The limitation to EEA jurisdictions would discriminate US transactions and leads to misallocations. We think that regulation here by including a jurisdiction threshold is redundant. We very much welcome the disclosure of loan tapes that include all important information regarding the jurisdiction of the underlying loans. It should be up to the investor to choose the country in which it wants to be exposed to risk. Chinese auto loans originated by VW, for example, might not expose the investor to higher risk than Greek auto loans.

For us it is important to highlight that the reporting can be improved for securitisations. Portfolio reporting should be standardised and more transparent. Further, it should be accessible from one data warehouse (ideally Bloomberg)."
In general, we would expect larger market-spread differentiation between qualifying and non-qualifying securitizations due to different capital requirements. The difference between the treatment of securitizations in Basel III and Solvency II leads to adverse market consequences, because insurance companies are crowded out of the securitization market due to higher capital requirements imposed on them compared to banks. The qualifying securitization framework should lead to favourable capital treatment within both regulation frameworks - Solvency and Basel.

In our view it is important that some asset classes are not mis-specified as non-qualifying securitizations and unfairly discriminated. We think that CLOs should be included in the simple standard and transparent securitization framework and enjoy better regulatory treatment. Performance with regard to default rates has been good over the past years and even throughout the credit crunch. These transactions do not deserve to be treated like CDOs, CDO squared or bespoke transactions, which have showed a disastrous performance since 2007. Currently the regulator aggregated CLOs to CDOs. It is worth to mention that many CDO securitizations do not exist anymore.

(please see graph 2 of the attached file)
The partition between qualifying and non-qualifying securitisation should be reasonable and take into account the different risk profile. The required capital charge should also consider the differences between securitisations and covered bonds, respectively German Pfandbriefe in particular. Covered bonds and German Pfandbriefe are of higher quality given the dual recourse and separate legal regulation. The securitisation process as such adds additional risks.

The definition of qualifying securitisation should be valid for all tranches of the transaction and not only for the senior class as under Solvency II. The realised impairments and losses for type I and type II assets do not justify jump in capital charges for the two assets types.
We could imagine that an expected loss rating methodology (like Moody’s) would provide an adequate indication of the riskiness (expected loss) of a rated tranche.
As outlined in the ECB & Bank of England paper (The case for a better functioning securitization market in the European Union / May 2014) simply looking at the price risk over the period of the financial crisis would be too shortsighted. Special circumstances like the breakdown of the Asset-Backed Commercial Paper (ABCP)-market which have led to the extreme market stress cannot simply be projected in the future. In our view a fundamental realized loss component of the underlying asset portfolio and the realized impairment of tranches should be included in the calibration as well. Further there should be consistency between the capital charge of the underlying assets and the securitization.

We support the results of the EBA paper that the overall capital charge of a securitization is higher than the capital charge of the underlying portfolio. Maybe the capital charge of a securitization should be a combination of underlying asset capital charge and rating of the tranche (in order to reflect the structural enhancement). For example lower Mezzanine tranches get a multiple above 1 on the underlying asset capital charge (to reflect the leverage) whereas higher mezzanine and senior tranches get a multiple of below 1 to the underlying asset capital charge (to reflect the credit enhancement). Please note that the example mentioned in the sentence before is purely for demonstration purposes only and a proper calibration is needed.
We could imagine that rating agencies apply a shadow rating", which addresses the rating without the country ceiling. Capital requirements for simple standard and transparent securitization should be linked to the respective shadow rating."
German Insurance Association (GDV)