Revised SMA could allow banks to ignore past op risk losses

The Basel Committee’s proposed revisions to the standardised measurement approach (SMA) for operational risk capital will give national regulators the freedom to let banks ignore past op risk losses when implementing the framework, Risk.net has learned.

If adopted as proposed, the revised approach would likely result in significantly lower than expected increases in operational risk capital versus the committee’s initial SMA proposal from March 2016 – which could have led to a 75% increase in Pillar I operational risk capital according to one industry survey. One European dealer expects the new methodology would produce capital requirements on a par with those it currently holds under the advanced measurement approach (AMA), while one US dealer says it could result in significantly lower capital.

“Simplicity won the day. This SMA doesn’t end up giving you something directly comparable across banks because of this national discretion. Across the piece, you are going to have winners and losers – that happens when you apply a single formula to all banks,” says Dimitris Bartzilas, head of operational risk management capital at Credit Suisse.

The Basel Committee declined to comment.

Coupled with the option to allow for a lengthy phase-in of the capital requirement over a seven-year period, the proposals open the door to a wide disparity in the approach’s application across jurisdictions – seemingly negating one of Basel’s stated aims in switching to the SMA of achieving greater international comparability between op risk capital levels.

“The whole basis for the SMA was a lack of comparability internationally and the need for stable numbers. If you come up with a framework that moves away from comparability by allowing national discretion you will be left with the exact same problem you have right now – which is huge jurisdictional differences,” says Evan Sekeris, a partner in the financial services practice at Oliver Wyman.

The proposals were detailed in a May 19 letter, seen by Risk.net, from outgoing committee chairman Stefan Ingves to the group of governors and heads of supervision (GHOS) who oversee Basel’s work, ahead of a meeting of the secretariat in Lulea, Sweden, which took place earlier this week. Multiple sources suggest no deal was reached at this week’s meeting, however – meaning the framework could be revised yet again before the package is ready to be signed off by GHOS.

“At national discretion, jurisdictions may exclude the loss history from the calculation methodology, though all banks will be required to disclose their historical operational losses,” writes Ingves.

The extended lead-in time makes some uneasy, as it would allow banks to hold significantly lower amounts of operational risk capital deep into the next decade. “It is an eternity from now to 2027,” says a source at a prudential regulator.

Income measure

The proposed framework would rely on a simple accounting measurement of income – dubbed the business indicator (BI) – to divide firms into three different size buckets. A separate BI multiplier would be applied to each bucket to produce the capital charge. From the smallest size bucket to the largest these would be set at 12%, 15% and 18%, respectively.  The March 2016 SMA proposal featured five BI buckets, with coefficients set at 11%, 15%, 19%, 23% and 29%.

Bartzilas says the switch was achieved by merging the second, third and fourth buckets from the March 2016 proposal into one. Bucket 1 will contain banks with a BI range of €0–1bn; bucket 2 €1–30bn; and bucket 3 €30bn and above.

Shrinking the coefficients as proposed would produce lower capital requirements than those predicted under the 2016 SMA.

“That 18% maximum would create a situation where all US banks would end up with lower capital. Right now the loss distribution approach [used in the US AMA] produces a number that is that beyond 18%,” says the head of operational risk measurement at a US bank.

The 2016 SMA also required banks to calculate a loss component – an average of bank-specific operational risk losses over the previous 10 years – and plug this measurement into a regulatory-defined formula alongside the BI component to produce the ultimate SMA charge. This loss component was intended to enhance the framework’s risk sensitivity and provide incentives for banks to improve their op risk management.

Across the piece, you are going to have winners and losers – that happens when you apply a single formula to all banks

Dimitris Bartzilas, Credit Suisse

However, the May 19 letter indicates that national regulators would be allowed to exclude the loss component from the calculation – and with it a link to a bank’s operational risk history. Banks would still be expected to disclose their historical op risk losses, though. 

Dealers say this makes the SMA little more than a progressive tax linked to a bank’s size. “It is clear that there will be no risk sensitivity in the new approach; rather capital will be purely a factor of size,” says Lav Bernfest, head of operational risk control at Erste Bank.

Others refute the idea the 2016 SMA was ever truly risk sensitive: “Making that loss component discretionary certainly makes it more difficult to make things comparable, but to my mind it doesn’t really change the risk sensitivity. The fact is this isn’t a risk-sensitive regime, it’s a simple regime. There is no more or less risk sensitivity than in the AMA model today, which is completely backward-looking,” says the head of operational risk at a US bank.

The May 19 letter also introduces new transitional measures that would save banks from having to apply the set BI coefficients until 2027. In 2021, for example, national supervisors would have discretion to set the coefficients to 6%, 9% and 12% respectively.

As the transitional measures are voluntary, they could also lead to different capital requirements being imposed on banks of similar size but in different jurisdictions.  

“My expectation is there will be some disparity. No way is everyone going to go to the low end,” says the US head of op risk measurement.

Sekeris adds “there is no chance on earth” that the proposed calibration of the BI and coefficients today would be appropriate for banks in 10 years’ time – when the rule takes full effect – as the relationship between institutions, their income measures, and exposure to operational risk would have changed.

“By the time it goes out, it will be irrelevant because people will have moved on; the understanding of the risk would have evolved,” he says.

The US head of op risk measurement, however, argues there are benefits to a lower capital floor. “Having a calibration that is closer to a floor is very important to help people to have an incentive to measure the risk. If you have something draconianly high, the way it is now, you kill the incentive to actually want to measure it,” he says.

Springboard

Far from ending the practice of op risk modelling, the stripped-back SMA could yet herald its renaissance, some argue.

“This is an opportunity to focus on models that explain risk better than the AMA. If we can use the opportunity of the SMA to take resources from the AMA function to get models that better understand our operational risk, that would be a positive outcome,” says Bartzilas. 

National watchdogs are also likely to continue to demand model-produced figures from banks under their supervision in spite of their exclusion from Pillar I. 

“I expect that the regulatory focus will shift to Pillar II, where banks will face a more robust challenge from regulators, especially in the EU where the Supervisory Review and Evaluation Process is increasing in complexity every year,” says Erste’s Bernfest.

Market participants speculate that the May 19 letter represents a political compromise between different members of the Basel Committee, one that was needed to keep the Basel III regulatory package on track. Sekeris says European regulators worked to reverse-engineer the numbers their banks already produce under the AMA to ensure the SMA would not result in a material capital increase.

Yet dealers say this kind of deal was necessary. “A compromise had to be found because the SMA could not possibly have been implemented the way it was initially designed. It would not have been possible to force banks in some jurisdictions to pull up their operational capital by 40% or more; it was nonsensical,” says an operational risk head at a European bank.  

A formal announcement of the revised SMA, together with the other components of the Basel III package, is unlikely before it receives approval from the GHOS. Assent may be given at the time of the next G20 meeting in Hamburg in July, though this is far from certain.

Additional reporting by Tom Osborn