Members of The NeuGroup’s T30-2 measure counterparty risk in a number ways.
A least a few companies’ major bank counterparty made big news recently with a more than $2bn trading loss. But while that bank’s, (OK, JPMorgan’s), trading losses are big, they’re likely not enough to put it out of business – it reportedly has a $360bn derivative trading portfolio that has earned billions over the years. So the ISDAs et al likely haven’t changed all that much. Still… anything can happen, so it’s always prudent to check on all the methods used to check on what risks the company’s counterparties present.
A recent survey of The NeuGroup’s Treasurers’ Group of 30-2 (meeting this week in New York City) revealed that companies are using a variety of ways to measure risk. According to respondents, the top two metrics used to measure counterparty risk are credit ratings (94.4 percent), and CDS spreads (83.3 percent). Sixty-one percent of respondents have CSAs for OTC derivative transactions, and most use symmetrical triggers (44.4 percent). Meanwhile one third of respondents include the impacts of collateral management in treasury reporting, while another 22.2 percent plan to do so in the future.
A little over 55 percent of respondents use notional exposure outstanding. Interestingly, the International Swaps and Derivatives Association (ISDA) frowns on using this measure. “Notional amounts are only loosely related to risk,” ISDA has said. “For most OTC derivatives, cash flow obligations are normally a small percent of notional amounts and so are mark-to-market exposures. Further, netting of obligations under a master agreement and collateralization of exposures reduces credit exposures to less than one percent of notional amount.” It should be noted that more than half of T30-2 survey respondents use mark-to-market exposure to measure risk as well.
More pertinent to recent events, a few members of the group said they factor in stock price movement and press reports. No doubt recent problems at JPMorgan have those model bells ringing.