NEW YORK – Revolutions in financial fortunes over the past year provide as many lessons for the implementation of Basel II models as they do for transparency, according to Sheila Bair, chair of the Federal Deposit Insurance Corporation, in a speech to the Global Association of Risk Professionals.
Bair spoke about the modelling consequences of shifting ratings for products such as structured investment vehicles, monoline insurers, municipal bond auctions and mortgage-backed securities.
Bair said: “Reflecting this belief in models and ratings, many were impatient with a go-slow approach to Basel II. They did not want capital safeguards, and they did not want the US leverage ratio. Today the financial world looks and feels vastly different.”
Bair said that regulators have increased risk exposure through Basel II’s advanced measurement approach, by lowering the risk weight on many of these collateralised debt obligation securities by 7% or less.
“The advanced approaches in general represent a heavy bet on the accuracy of models and quantitative risk metrics. Based on specific assumptions about loss correlations, the framework promises heavy reductions in capital requirements for virtually any credit class with a favourable loss history. And this can encourage banks to lever up far more than they already have in a competitive dynamic to boost their return on equity.
“So, here’s the bottom line: we have a major lack of transparency in structured finance. And we have regulatory policies, including Basel II, that may have the unintended consequence of encouraging some banks to bet heavily on non-transparent ratings and models,” said Bair.
She highlighted the importance of addressing these issues as part of a careful evaluation of the Basel II securitisation framework, emphasising that banks will need incentives to use the information rather than just making it more readily available through transparency.
“Regulatory policies should not encourage banks to ignore traditional credit analysis in favour of placing financial bets whose success depends on the accuracy of any particular rating, or particular model,” she said.
The week on Risk.net,October 14-20, 2016Receive this by email