Three Japanese megabanks are expected to start reflecting the market value of counterparty risk in their derivatives portfolios for the first time in the next round of financial results, a change in the banks’ pricing methodologies that could result in sizable revaluation losses.
Experts believe the losses could be on a similar scale to the $712 million hit to Standard Chartered in 2016, which followed revisions to the way the bank reports its credit valuation adjustment (CVA).
Until now Mitsubishi UFJ Financial Group, Mizuho, and Sumitomo Mitsui Banking Corporation – like all Japanese banks – have had to hold capital against CVA, but did not have to report its fair value in accounting statements. Nomura is an exception because it already reports CVA.
Introducing the measures will require the three banks to record a one-off downward adjustment to their profit and loss statement. It is unclear exactly how large the losses will be given the lack of clarity on the methodology the banks use, and which contracts the CVA will initially apply to.
But considering the size of the banks’ swap business with domestic and foreign corporate clients, which tends to be uncollateralised, any revaluations – if applied consistently across the banks’ derivatives portfolios – are likely to be significant, says Satoshi Kumeta, head of XVA in fixed income, currencies and commodities at Daiwa Securities in Tokyo.
“I cannot say how much [P&L impact] but I know the megabanks have huge uncollateralised derivatives portfolios,” he says.
Japanese banks don’t publicly disclose information on uncollateralised derivatives, but dealers estimate the range is from tens of billions of dollars at the lower end, to hundreds of billions at the upper end.
“Standard Chartered-like losses certainly gels with what I have heard,” says Michael Reeves, a Tokyo-based financial markets consultant. “Ultimately your position is going to look substantially worse if you traditionally don’t include a metric that has a negative impact. I would be shocked if that were not the case.”
A London-based XVA consultant says the size of the revaluation is dependent on certain factors that are not evident from reading financial disclosures, such as the tenor of the book and how the bank plans to hedge CVA.
“My guess is that if they do it properly and in a market standard way the losses would be bigger [than Standard Chartered] for the megabanks, on the basis that it could be driven by very long dated, uncollateralised exposures,” says the consultant. “If not then I certainly don’t think it will be too much smaller.”
Also in 2016, ANZ Banking Group took a $120 million hit on CVA revaluation.
CVA is calculated using three inputs: loss given default (LGD), the percentage of an exposure an entity would expect to lose if its counterparty defaulted; probability of default of the counterparty; and estimated exposure at given points in time. Historical LGDs are based on observed recovery rates on comparable counterparties, while market-implied LGDs are derived from credit default swap spreads. The two approaches can produce vastly different CVA numbers, with market-implied being larger.
The adoption of accounting CVA in Japan stems from discussions between the banks, regulators and industry bodies as part of a working group convened by the Japanese Bankers Association in 2017. The working group agreed that banks should base CVA calculations on market-implied rather than historical LGDs.
The group did not set a deadline for adoption of accounting CVA in its official report, released in June 2017, but during the discussions the Japanese financial regulator, the JFSA, is said to have made it clear that it expected CVA to be reflected in the banks’ financial results. The banks are thought to have discussed starting such reporting from March 2019.
Standard Chartered-like losses certainly gels with what I have heard
The individual that represented the JFSA in the working group has since left their role. Sources say it is not clear whether the personnel change will affect the regulator’s expectations of the accounting CVA rollout, or its timetable. A JFSA spokesperson says the personnel change will have no effect on policy, adding that the regulator “expects that the industry and each of the firms will deliver steady progress toward the goal set out in the report”.
An XVA expert at an international accounting firm in Tokyo notes there is a strong regulatory incentive for the banks to use market-based LGD, rather than the historical method, in their accounting statements.
To be able to apply the standardised approach for calculating regulatory CVA under new rules known as the Fundamental Review of the Trading Book, banks must have a CVA framework in place that includes the use of market-based CVA calculations. If banks are unable to use the standardised approach, known as SA-CVA, they will have to fall back on to the basic approach, which could lead to capital charges up to 10 times more punitive for some regional banks.
“It is very important for the banks to maintain consistency between regulatory CVA and accounting CVA,” says the Tokyo-based XVA expert. “They are going to eventually need market-based CVA before the enforcement of FRTB SA-CVA.”
The CVA losses might not be easy to identify in banks’ accounting statements, however, especially if banks opt to phase in its introduction, which could start with plain-vanilla derivatives before rolling out to more exotic products. This could add another layer of uncertainty to estimates of the initial P&L impact. Together with the leeway that exists around how CVA can be presented in financial statements, it could allow the banks to disguise the full impact of the change, notes Daiwa’s Kumeta.
“I don’t think it is usual to show XVA in financial statements explicitly in Japan, thus it will be probably difficult to distinguish them from other P&L items in their statements unless they add some notes,” he says. “The Japanese Bankers Association recommended a phased-in approach, so there is also a possibility that they will introduce it step-by-step, which could make the total P&L impact unclear.”
Editing by Alex Krohn