Risk transfer regulatory, structuring considerations discussed

Representatives from Reed Smith and StormHarbour Securities recently provided an introduction to capital relief trades, focusing on regulatory and structuring considerations, during a live webinar hosted by SCI (view the webinar here). This Q&A article highlights the main talking points from the session. For a broader and more in-depth exploration of the risk transfer space, attend SCI's Regulatory Capital Relief Trades Seminar on 3 December.

Q: With respect to the regulatory framework for capital relief trades, how has the sector evolved?
Claude Brown, partner at Reed Smith: To a certain extent, the regulatory environment for capital relief trades pre-dates Basel 1. The Bank of England was the first central bank to formulate rules around the transfer of risk, related primarily to participations. It also issued generic rules about guarantees.

Those rules developed along with the evolution of the debt trading market in London, principally around Russian Vnesheconombank debt. Then, Basel 1 introduced the concept of credit conversion factors, which dealt with off-balance sheet assets, such as guarantees and stand-by letters of credit. Around this time, we also saw the credit derivatives market begin to emerge.

Some of the first risk transfer trades - such as the ROSE Funding deal for NatWest - used participations. Several others used credit-linked note structures, such as the SBC Glacier Finance transaction and Credit Suisse's Triangle Funding.

But the real breakthrough into what would subsequently be considered as synthetic securitisation was JPMorgan's BISTRO transaction. This deal was innovative in that it allowed the portfolio to be larger than the amount of notes issued.

It also managed to overcome the problem that had beset some of the earlier trades, where the credit rating of the originator acted as a cap because the CLN or participation proceeds sat with the originator. BISTRO used the capital for two purposes: to collateralise the repayment obligations of the SPV under the notes; and to provide collateral for a credit default swap, under which it sold protection to JPMorgan on a portfolio of underlying securities.

There were some limitations, however. The transaction didn't qualify for full capital recognition and the collateral had to run to term before credit events could be paid out. That issue was rectified in later deals, such as BNP Paribas' OLAN, where more flexible means of collateralising deals were used.

Concurrent with these developments was the emergence of the regulatory framework, which borrowed heavily from the Bank of England concepts and was reflected in the various Basel iterations. In Europe, the Banking Coordination Directive dealt with some risk transfer features and what needed to be achieved for regulatory capital purposes. More recently, the Capital Requirements Directive and the CRR came into force.

Obviously the US Fed and the OCC have their own rules for US-regulated banks. APRA in Australia has its own set of rules for risk transfer, which makes it challenging to structure a trade successfully. Even under the CRR, EU regulators have differing interpretations.

These are the primary regions where banks have been involved in CRTs, but obviously other regulators have their own rules too.

Q: In terms of where the market stands today, what is the general attitude among regulators towards CRT transactions?
CB: It varies widely. There is recognition that it's prudent for banks to try and manage their risk and to do that by diversifying it and in some cases transferring it to those who are able to manage it.

This is recognised on two levels, one of which is the economic benefit, whereby the capacity of banks to stimulate growth is increased. The other is that prudent risk management holds that there is a proper place for risk transfer.

Balanced against this is concern over whether risk can be monitored to minimise any adverse impact arising from systemic risk. That produces a spectrum of responses from regulators, from those that are generally against risk transfer to those that recognise it has a useful role within the financial system.

Robert Bradbury, md at StormHarbour Securities: The attitude of local regulators to risk transfer trades in general is also somewhat driven by local history and legacy issues. In some jurisdictions, regulators have had to deal with some transactions that would not be compliant with the modern view of risk transfer, for example.

Q: How are capital relief trades typically structured?
RB: The most common assets to be referenced in the market include SME, corporate and shipping loans. This is due to a variety of reasons, the first of which is that data is more readily available and of better quality for these assets.

From an efficiency perspective, they hold favourable risk weights relative to expected loss. There is also a high degree of familiarity with the assets among investors. Finally, rating agencies are able to find more historical data on those assets.

Less common assets include PFI, consumer loans, retail unsecured and real estate. There are a few reasons for this, the main one being the lower risk weight and higher expected loss per unit.

Data is also less available, which makes it harder for investors to do due diligence. Additionally, there are legal issues regarding disclosure of personal information in some jurisdictions, which can complicate the process.

Certain exposures are categorised under the IRB slotting approach, which makes the process more difficult for originating banks.

The typical structure for an IRB bank is based on the supervisory formula, rather than the ratings-based approach. This necessitates retaining the most senior part of the capital structure and placing the mezzanine tranches or possibly equity.

In terms of putting a trade together, the first step is to identify the portfolio and all the resulting data requirements. Already having a cash securitisation programme can make the task easier. But if an issuing bank is trying to gain protection on an asset class that it wants to strategically exit, identifying a portfolio can be tricky.

The next step is to calculate risk weights and KIRB, which is the main capital metric in securitisations and is a function of loss given default, probability of default, maturity and asset type.

Often the easiest way to execute a trade is bilaterally, where the bank faces a single counterparty under a bespoke contract - with no ISINs, SPVs or transfers involved, but a CDS referencing a tranche of a synthetic portfolio. However, the more common execution is via an SPV, which provides protection to the issuing bank using either a financial guarantee or a CDS contract. Generally, the CDS is held on a mark-to-market basis, while a guarantee - while being more difficult to structure - is often eligible for accrual accounting.

The investor effectively purchases a CLN in exchange for an upfront sum of cash equal to the size of the hedged tranche. The CLN is linked to the underlying hedged tranche.

Two types of structures are most common: the simplest is a fixed-leg payment, where the floating pay-outs are linked to the contract; arguably more complicated is where a synthetic waterfall is referenced - similar to a cash ABS transaction - that is linked to the spread of the portfolio. A virtual senior tranche is used to compute the deduction from that spread and any excess spread, subject to a cap, is passed down to the mezz or equity investor.

Which method to adopt is fairly jurisdiction- and asset-dependent. A typical cash ABS investor may prefer a synthetic waterfall, for example, because it's easier to reconcile with their existing holdings. However, these structures can impact the way in which significant risk transfer is demonstrated.

Proceeds from the CLNs are held in a cash account. Although it's possible to execute unfunded trades, overwhelmingly trades are funded because it reduces counterparty risk, reduces the risk weight of the tranche and helps with the designation of risk transfer.

One of the requirements for efficient risk transfer is for losses to be paid out promptly. In order to satisfy this requirement, when a credit event has occurred, the cash account is stepped down directly.

The mechanism through which this is done can be LGD-based, fixed upfront or 100% upfront. Following a pay-out, there is a recovery period, which again is asset-dependent and undertaken in accordance with the bank's normal workout process. That process concludes when there are no more recoveries or at the backstop date when the asset is resolved via an auction sale and the seller often has a 'last look' for physical delivery to protect against zero bids.

The bank retains any tranche junior to the equity, e.g. 0%-0.5% or 0%-1%, plus the senior tranche. It also has to comply with the 5% risk retention requirement under CRR, commonly by retaining 5% of the note and the other tranches or 5% via a random selection of exposures.

The cost/benefit analysis is fairly unique to each bank, but predominantly they look at three metrics, the first of which is a point-in-time metric that assumes zero loss and a one-year time horizon. The second is a through-the-cycle method, which involves checking the efficiency at multiple points throughout the life of the trade and that necessitates cashflow modelling and RWA modelling.

The more complicated method is a weighted average cost of capital calculation, where you discount a number of risky cashflows at a risky discount rate computed according to how the bank would look at this trade over a period of time. We've found anecdotally that this is often done using a comparable cost of capital, such as issuing equity.

Q: If an investor sells protection on the 1%-10% tranche and the bank retains the first 1%, does this count as greater than 5% risk retention by the bank?
RB: No.

Q: Can capital relief be achieved on a portfolio under the standardised approach?
CB: Because the standardised approach is a relatively blunt instrument, the efficiencies are harder to realise. Besides, a bank that doesn't have sufficient modelling capability to move onto IRB is likely to struggle with marshalling the data for a risk transfer trade.

Q: What are the main requirements for achieving significant risk transfer?
RB: Significant risk transfer is a topic of considerable focus, given that a number of trades were done in the past that wouldn't necessarily comply with the current guidelines. The SRT regulations come from the CRR and we have the EBA guidelines as well, which the European Commission will hopefully now implement.

Arguably the most important point under SRT is the high cost of credit protection: you have to show that the rate you're paying for your protection in a zero loss, expected loss and stressed scenario does not undermine the level of credit protection. This stipulation is designed to prevent transactions being done where payouts exceed the size of the tranche and implicitly caps the 'fair' amount you can spend. Anecdotally, some regulators appear to have a more stringent view on the high cost of credit protection than others.

You also have to demonstrate the robustness of internal models, including the stability of the supervisory formula over time. The documentation should obviously reflect the economic substance of the transaction and there should be no implicit or explicit support. The securities issued should not represent payment obligations of the originator.

Additionally, structural features such as clean-up calls and replenishment rights need to be assessed to ensure compliance with SRT. Any maturity mismatch in the portfolio also needs to be taken into account.

The bank needs to have policies and procedures in place around risk management and self-assessment. Finally, the form of the credit derivative used can't undermine SRT.

Many of these requirements are becoming increasingly standardised, helped by the EBA guidelines and the fact that these transactions are becoming more common.

Q: What are the drivers behind why originators and investors participate in capital relief trades?
RB: From an originator perspective, the main drivers are risk reduction and increasing the stability of RWAs. But it's important to distinguish between a point-in-time static short-dated transaction and a longer-dated through-the-cycle transaction. Static point-in-time trades are more appropriate for run-off books or specific situations, while longer-dated trades are more appropriate for live, revolving portfolios.

The three- to five-year tenor is common, where originators are trying to achieve not only immediate RWA benefits, but also through-the-cycle protection. Revolving transactions allow for stability of capital relief: to the extent that the replenishment provisions allow, replacing loans when they amortise maintains the efficiency of the trade and mitigates credit risk migration over time.

There is a balance to strike with tenors. Hedging against migration is good, but you have to avoid any implication of implicit support through asset selection. You also have to pay for increased effective duration, which can in turn limit the investor universe.

Indeed, cost of capital often acts as a ceiling for the execution of these trades.

Diversification is the main motivation from an investor perspective. CRTs allow them to invest in assets that aren't normally available outside of banks and to take a macro view on a given sector or country, or to leverage their micro view on concentrated portfolios.

In some cases, the opposite is true and the investor is effectively taking a view on the bank's whole portfolio and underwriting process.

Q: New risk retention rules have focused attention on alignment of interest within the capital relief space. How are these concerns being addressed?
CB: One of the main tests for alignment of interest is significant risk transfer, which ensures that the bank has genuinely exported a significant amount of its risk. However, the risk retention rules are somewhat of an opposing force, necessitating banks to keep skin in the game. Either way, clearly investors want to ensure that they're not exposed to the moral hazard of a bank's indifference as to how the portfolio performs.

The risk retention rules are probably the most challenging because they impose obligations on investors, assuming they're regulated. It's their responsibility to ensure that risk retention is observed, diligenced and monitored for the life of the deal. Investors have to ensure they're comfortable, even though the means of compliance sits with the originating bank.

The retained sliver of the equity tranche is a popular mechanism for compliance. We also see the vertical tranche across the mezzanine from time to time, which makes it easier to finance the trade. The random 5% retention is only really appropriate for a very granular pool.

RB: Alignment of interest came out alongside the requirements for transparency and disclosure of all relevant information. Data requirements can be quite heavy and investors need to be able to confirm that they've received all materially relevant data. Different banks assess this in various ways.

Q: In terms of executing capital relief trades, what are the main issues that participants should be aware of?
RB: Obviously there is a swath of considerations when putting a risk transfer trade together. Structural considerations include defining your protected tranche: it needs to be an equity or junior mezzanine tranche, but the attachment and detachment points are functions of the types of assets, the budget of the issuing bank in terms of premium spend, the amount of benefit they want to receive, investor requirements and the jurisdiction.

The format is driven by a bank's tolerance of mark-to-market volatility and desired speed of execution, as well as investor requirements. Some accounts can't invest in CDS or financial guarantees, for example.

Asset type, granularity, seasoning, tenor and spread are all important portfolio characteristics. For more concentrated portfolios, the identity of some or all of the borrowers may need to be disclosed to investors, which can complicate the process. Data availability, loss history, provisioning levels and future anticipated asset availability all influence the structure, target investors and pricing level.

Target maturity will be based on the assets and the bank's goals. The transaction may be static or revolving, in which case replenishment and eligibility criteria need detailed consideration.

Some transactions may require ratings for the senior tranches, in which case rating agency requirements need to be met.

Collateral can be held by a third party (for lower rated banks) or the issuing bank itself (for higher rated banks).

Amortisation of the portfolio dramatically affects a trade's WAL and how premium is calculated. Amortisation is either pro-rata or sequential, based on repayments and prepayments. Different jurisdictions have different views on how this affects SRT.

Credit events normally comprise bankruptcy, failure to pay and restructuring. Documentation often includes specific features to meet local regulatory or legal requirements.

Placement can be done bilaterally or through a private auction with a number of parties. Auctions can improve pricing tension, but there are scenarios in which a bank doesn't necessarily want to advertise the existence of a portfolio. There are also cases where certain high loss or exotic portfolios may not attract enough interest to warrant an auction and so dealing with one or two parties with a known appetite for that kind of risk makes more sense.

Some investors demand more significant alignment of interest than 5%.

Q: What type of analysis should be considered when an investor requests a risk alignment of, for example, 20%?
RB: One consideration should be if this impacts whether the trade can still be considered as transferring significant risk. This will partly depend on how the alignment is performed.

Other considerations could include the materiality of remaining potential losses and the increase in the amount of assets effectively being referenced. Another point could be the effect on cost efficiency after such retention for the issuing bank.

CB: Once you move above 5% risk retention, you have to be mindful that under some regimes this will have a direct impact on whether you meet the SRT test.

There are a number of measures in the CRR that can be used to demonstrate SRT. One of these is where there are no mezzanine tranches in a securitisation and the originator does not hold more than 20% of the exposure values of the securitisation positions that would be deducted from its CET1 and the originator can demonstrate that the exposure values of the positions that would be deducted from CET1 exceeds a reasoned estimate of the expected loss of the positions by a substantial margin.

So, at the 20% alignment level, there is a risk that SRT has not been achieved. That said, it doesn't mean that below 20% it will be achieved, because the originator may fail limb two of the SRT test or be using one of the other three SRT methods.

Q: Do regulators have a preference with regard to sequential versus pro-rata structures?
RB: It's jurisdiction-dependent, but there are two countervailing considerations. First is that sequential amortisation preserves the value of the tranche until maturity (in the absence of losses), meaning you have the largest amount of credit protection available for a single event for the longest period of time, but this dramatically increases the premium. In comparison, the benefit of a pro-rata structure is that the protection tracks what is occurring to the portfolio - although this can be perceived as an issue for concentrated portfolios near the end of their lives because the tranche may no longer be large enough to absorb a credit event.

CB: It also depends on the sophistication of the regulator. The CRR applies to 28 regulators with varying degrees of experience of these products.

Q: In which situations would a transaction be rated by a rating agency?
RB: Other than for usage of the ratings-based approach, a rating may help with demonstration of SRT, particularly for portfolios or transactions with unusual characteristics.

CB: Ratings are sometimes requested by an investor because of internal or regulatory requirements, or because it helps them monitor the transaction.

Q: Looking ahead, how is the risk transfer sector expected to evolve?
RB: We're in a good period for this type of transaction. The market has evolved significantly in recent years and the level of regulatory and governmental involvement is unprecedented. Awareness that these trades exist and can serve to help the wider economy is increasing.

Having a single supervisor in Europe is expected to help the market in terms of implementation of regulation, such as the SEC-IRBA, SEC-ERBA and the simplified versions further down the hierarchy of the new approaches. This provides clarity as to where the market is heading and the process should therefore be more transparent in the future.

Harmonisation of RWAs and the implementation of floors may result in higher RWAs, which would make CRTs more attractive.

A number of new players are entering the market, including development banks and the SME Initiative, which will help the sector. As more trades are executed, smaller banks will inevitably follow the larger ones. Proof of concept of these deals is broadening and we can expect to see increased use of the technology in Eastern Europe, for example, and more asset classes being referenced.

The STS criteria may well be extended to cover synthetic securitisations, which would be a positive development, but in any case cover cash securitisations that may be used for risk transfer purposes. Indeed, cash structures may become more common, due to their increased flexibility in terms of the leverage ratio and so on.

Finally, the secondary market continues to develop and this will improve liquidity and pricing for some issuers, as people become more familiar with them. Standardisation of the product is a long way off, but that's the direction in which the market is headed.

Q: How is the implementation of the revised securitisation framework and the debate on simple, transparent and standardised securitisation impacting the CRT space?
CB: The incentives for sophisticated banks to move from the standardised approach to IRB will enable them to engage in these trades more easily. CRTs lend themselves better to IRB.

There is also a virtuous circle in that if you're providing more data and information on an ongoing basis to deal with risk retention requirements, it encourages banks to acquire data on the portfolios more readily, which allows them to do more issuance. Scoping potential portfolios becomes easier over time.

Having said that, the STS proposals in their current form clearly set themselves against synthetic securitisation and anything that involves less than a complete transfer of the asset pool. Of course, you can execute cash-based transfers that would comply.

Regulators and policymakers recognise that risk transfer and securitisation play a role in promoting economic growth. The thawing of hostilities towards securitisation is positive, but we're yet to see the full effects from the new regulatory framework play out, so the market is likely to remain challenging until 2018. I expect a slow and steady growth of CRTs, as well as measured improvements in the technology and the range of issuance.

SCI's Regulatory Capital Relief Trades Seminar is being held on 3 December at Reed Smith's offices at 20 Primrose Street, London. The conference programme consists of panel debates covering risk transfer regulatory treatment, structuring considerations, pricing trends, opportunities and relative value, and capital alternatives.

Speakers include representatives from: Apollo Global Management; Caplantic; Chenavari; Chorus Capital; Christofferson Robb; Citi; Clifford Chance; Elanus Capital; European Investment Fund; Lloyds; Mariner Capital; Natixis; Nomura; Reed Smith; StormHarbour; UniCredit; and West Face Capital.

Click here to register.