Singapore’s largest lenders are looking at using their own models to calculate capital for foreign exchange trading desks, after recent changes to market risk rules alleviated concerns among Asia’s banks over the uncertain capital impact of the new regime. However, adoption of the internal models approach, or IMA, will not be straightforward, banks warn.
“We are going for IMA,” says Chew Chee Keong, UOB’s managing director for market risk. “We can save capital if we go with IMA for our foreign exchange trading desks.”
UOB estimates that shifting to internal models could reduce the bank’s market risk capital requirements by up to one-third compared with staying on the regulator-set standardised approach.
When the original market risk rules, known as the Fundamental Review of the Trading Book, were released in 2016, Asian dealers said the framework would hike capital requirements for less liquid markets such as theirs, largely because the data threshold for internally modelling risk factors was too high. Banks in other regional markets, such as South Africa, echoed these fears.
Even large US lenders had initial reservations about the IMA, with Wells Fargo committing to using the standardised approach.
Global standard-setters tweaked the framework in January to soften the rules around the observability of hard-to-model risk factors, which is a particular problem for less liquid markets in Asia-Pacific. The revised FRTB offers banks fresh incentives to push for internal models on certain product lines, and may prompt some Singaporean banks to extend their use to interest rate derivatives.
“For the markets side, we are exploring and assessing IMA for forex and linear interest rates desks,” says OCBC’s head of global treasury business management, Frederick Shen.
Singapore’s largest lender, DBS, is more equivocal. Brian Lo, group head of market and liquidity risk at DBS, says the bank is “not ruling out” pursuing IMA on a desk-by-desk basis.
Forex trading accounts for the largest portion of the derivatives book among Singapore’s banks. The city-state is the largest forex centre in Asia-Pacific, with average daily market turnover of $517 billion, third behind the UK and the US, according to the Bank for International Settlements’ Triennial Survey in 2016.
Singapore’s banks are planning to use IMA for their structural forex hedging. This type of exposure arises from forex fluctuations between a bank’s reporting currency and its holdings in subsidiaries operating in different jurisdictions. The risk sits in dealers’ banking books, as opposed to their trading books – but under FRTB, instruments that carry forex risk are subject to market risk capital requirements regardless of whether they are held in the trading book or banking book.
DBS reported a total of S$18.2 billion (US$13.5 billion) in structural currency exposures in the year to December 2017, while OCBC had S$20.7 billion and UOB’s stood at S$12.9 billion, according to their 2017 annual reports, the most recently available. The unhedged positions stood at S$5.8 billion for DBS, S$14.7 billion for OCBC and S$9.6 billion for UOB. The lenders are yet to release their 2018 annual reports.
Typically, banks use forex derivatives to hedge this structural forex risk. The hedging activity benefits from exemptions from market risk capital requirements – but banks must first seek their host regulator’s approval and meet certain conditions.
More broadly, any Singaporean bank intending to use IMA will need to gain approval from the Monetary Authority of Singapore and, for the first few years at least, benchmark its IMA results against the standardised model. Banks are expecting to learn details from the MAS on the city-state’s plans for FRTB and guidelines to define their trading desk structure.
The most capital-efficient structure for large banks, believes KokWah Lok, head of market risk client services for Asia-Pacific at tech vendor Murex, is to seek IMA across all forex trading desks.
“It may not make sense to split forex cash positions from the vanilla and exotic options desks,” he says. “Banks trade complex structures and then hedge them with spot and vanilla products. So if you split them into one desk governed under IMA and the other under the standardised approach, you don’t get the hedging benefits any more because you will set aside capital at different levels and without offsets.”
The price is right
The January revisions to FRTB are intended to help banks meet modelling thresholds in less liquid markets. The criteria for risk factors to qualify as modellable have been relaxed to permit four price observations in 90 days, or 100 observations in the previous 12 months. The 2016 version of FRTB permitted a maximum 30-day gap between observations for a risk factor to be deemed modellable. For products that trade on a seasonal basis, such as agricultural commodities, the new rules will aid banks hoping for IMA approval.
Singapore’s banks have an improved chance of meeting the IMA threshold for their forex desks because they have the historical data on prices and quotes, dealers say.
The January framework has also expanded the number of currency pairs that enjoy lower risk weights. Banks may now calculate foreign exchange risk in a currency pair by nominating a base currency, instead of the reporting currency. The revised approach enables banks to apply a lower risk weight to currency pairs that are triangulated from other, liquid currency pairs.
For example, a Singapore bank trading a liquid currency pair involving the US dollar and an emerging market currency should not then have to calculate the risk of the emerging market currency against their reporting currency – the Singapore dollar – because the dollar/Singapore dollar rate is highly liquid.
However, hurdles remain, banks warn. Liquidity in some Asia-Pacific markets is variable, which leaves trading desks susceptible to falling in and out of IMA approval as markets wane in liquidity.
“People need to assess how volatile IMA qualification would stay for a particular desk, especially when they are confronted by a significant number of risk factors that might get trapped in the non-modellable category, on and off,” says DBS’s Lo.
Banks are looking at ways to overcome these problems, including by pooling data.
“IMA has certainly got limitations,” UOB’s Chew says. He highlights the modellability of complex or exotic products as a particular challenge.
“Overall, though, there is still an incentive for us to move to IMA for capital purposes and better risk management, and we can reflect the impact through pricing to clients too,” Chew adds.
Editing by Alex Krohn