Synthetic risk transfers: a trillion-dollar tool that needs to take the next step
Prasad Gollakota
Flowback risk is a relatively new regulatory term. It aims to capture the risk that a bank may be unable to roll a synthetic risk transfer transaction, resulting in previously protected risk-weighted assets once again attracting full capital treatment.
SRTs allow banks to transfer credit risk on selected portfolios, reducing RWAs and supporting lending without selling the underlying loans. The International Monetary Fund estimates that more than US$1trn of assets have been synthetically securitised since 2016, underlining how SRT has evolved from a niche technique into a central component of the capital management toolkit.
Given its scale, the market has come under increased regulatory scrutiny. Both the Bank for International Settlements and the IMF recognise SRT as a viable capital management tool, but highlight flowback (or rollover) risk, alongside concentration risk, as key vulnerabilities.
Their concerns are well made. SRT transactions are typically short dated, often three to five years, and amortise over time, with capital relief diminishing as the underlying portfolio runs off. As a result, banks seeking relief against new lending growth must continue to renew these transactions. In addition, they are distributed to a relatively concentrated pool of specialist investors, including private credit funds and hedge funds, with limited participation from the broader institutional base. This combination means the product’s continuity is implicitly reliant on a narrow segment of capital markets remaining open and willing.
Focusing just on rollover risk, however, misses the deeper point. It frames the issue as one of refinancing and market access, when in reality the IMF’s observation reflects something more fundamental about how SRTs are designed.
Back to basics
Bank capital management is a multi-dimensional challenge of optimising the cost and structure of capital and the level of RWAs across time. With an SRT a bank retains a portfolio of loans on its balance sheet but transfers a defined layer of credit risk to investors through a credit derivative or note structure, often via a special purpose vehicle.
Investors receive a premium reflecting the risk, while the bank obtains regulatory capital relief by reducing the RWAs associated with the protected portfolio for as long as the protection is in place. The transaction is therefore a contractual transfer of credit risk rather than a source of funding, and economically, it can serve as a lower-cost alternative to the counterfactual of raising an equivalent amount of capital.
The counterfactual is important, however, because capital quality is typically assessed along two dimensions: its degree of permanence or longevity in supporting the balance sheet, and its ability to absorb shocks when they arise, whether credit losses or funding stress.
A standard 10-year non-call five Tier 2 illustrates this well. Under stressed market conditions, we have seen instances where a Tier 2 bond was not called at its first call date. Under Basel rules, Tier 2 recognition amortises on a straight-line basis over the final five years of maturity. For a standard 10NC5 that remains outstanding, that means amortisation begins after the first call date. But this reduction does not reflect any change in its legal or economic loss-absorbing capacity. The full notional amount remains outstanding as a subordinated liability. It continues to rank below senior creditors, remains eligible for bail-in in resolution, and continues to absorb losses in a gone-concern scenario. In other words, it serves multiple functions within the capital structure.
The original intent behind the amortisation of Tier 2 reflected a forward-looking view of capital quality, encouraging banks to maintain a stable maturity profile and avoid cliff effects in capital ratios as instruments approach repayment. Amortisation is therefore a regulatory overlay on measurement, rather than a statement about loss-absorbing capacity.
SRT does not behave that way. While it does absorb losses within the protected tranche up to a pre-defined detachment point, it is typically structured with a three to five-year tenor and amortises over time. Even if that term aligns with the underlying assets, the protection rolls off, with no embedded mechanism to extend capital relief beyond the contractual maturity.
One-trick pony
SRT operates on a single dimension and is, in effect, an insurance contract, with no built-in mechanism to bridge the transition between normal conditions and stress. That simplicity is part of its appeal: it is precise, targeted and, in many cases, economically efficient.
This is where the IMF’s concern about rollover risk becomes clearer. In benign conditions, transactions can be rolled, with stable pricing and sufficient demand. But in periods of stress, new protection may become unavailable or uneconomic precisely when it is most needed, constraining a bank’s ability to sustain capital relief and support further lending, potentially forcing it to slow origination, raise alternative capital or accept lower capital headroom. Either way, any capital plan built on continued access to the SRT market may have to be revised.
A question of design
Should a core capital tool rely so heavily on continued access to external markets, with no embedded extension rights or capacity to bridge stress?
Other parts of the market have faced similar challenges and evolved accordingly. Catastrophe bonds, which share many structural similarities with funded SRTs, recognise that risk is not always resolved on a fixed timetable. They incorporate extension mechanisms that allow protection to continue when losses are still being determined, with investors compensated through pre-agreed step-ups, broadly analogous to the step-up or reset features embedded in capital securities, which compensate investors for extension risk in the event of non-call.
These structures explicitly acknowledge uncertainty and are designed around it.
The challenge, therefore, is whether SRT can be designed to function across a wider range of market conditions. Could it incorporate contingent extensions, longer tenors than the original portfolio’s weighted-average life to allow new assets to be added as the original pool runs off, or structures that permit maturity extension and top-ups as scheduled maturity approaches? Some transactions include replenishment features, but these remain confined to the original transaction maturity.
Such design changes would not transform SRT deals into capital, nor should they. But they could allow them to behave less like a contract that expires and more like a tool that adapts. This would come at a cost, as investors would be asked to price the additional uncertainty, and the economics would shift accordingly. Regulators also have a role in facilitating thinking around extension mechanics, including through pre-approval at inception.
Capital tools should not be static. They must evolve through challenge and iteration, with design, pricing and investor appetite adjusting in response, and banks and regulators also playing an active role in market development.
Prasad Gollakota is a former FIG banker and co-head of the global capital solutions group at UBS. He was later chief content and operating officer at edtech company xUnlocked and specialises in financial institutions, banking regulation, capital markets and complex capital and funding solutions.