PRA sense, EBA analysis February 27, 2013 at 5:42 pm
Lord Turner told the Commission on Banking Standards that new banks would be allowed to start up with core capital equal to as little as 4.5 per cent of their assets, adjusted for risk.
The Prudential Regulation Authority, which takes over bank supervision from the FSA in April, would give new lenders some time – perhaps three years – to get their core capital up to 7 per cent. Their management would not be allowed to pay bonuses or dividends until they hit the higher threshold.
This is of course entirely consistent with Basel 3: 4.5% is the minimum, and between 4.5 and 7 you can’t pay divys or bonuses. So the UK is using all the flexibility that Basel gives them.
What really brings this into focus, though, is a passage later in the same story:
Small banks and new entrants must use a residential mortgage risk weight of 35 per cent – the key input for capital requirements – but some UK banks that use models have cut that number to 5 per cent
In other words, that is a 4.5% capital ratio based on standardised RWAs, not on IRB RWAs. The smaller or newer guys are not getting absolutely lower capital, just a little compensation for not having IRB models (or the history to prove that they are well-calibrated). As the EBA points out, IRB models can be manipulated to produce rather low capital requirements, making a 7% (or 10%) ratio easy to achieve.
The ‘top down’ analysis conducted using the existing supervisory reporting data from 89 European banks across 16 countries confirms material differences between banks in the calculation of the Global Charge (GC) defined as the sum of RWAs (unexpected losses) and the expected losses (EL).
The analysis conducted so far suggests that:
- 50% of the differences in terms of GC between banks mainly stem from the approach for computing RWAs in use (standardised vs IRB) as well as from the composition of each bank’s loan portfolio. In other words, these are differences that relate to the structure of a bank’s balance sheet as well as to its reliance on the different regulatory approaches for assessing and measuring risks (referred in the report as A-type differences).
- The remaining 50% stem from the IRB risk parameters applied thus reflecting each bank’s specific portfolio and risk management practices.
I wonder how much of this IRB-driven variation would go away if there was a 15% RW floor on mortgages.
A lot of it would go away if they analysed mortgages vs consumer finance at all (or SME vs large corporate). They didn’t get the data to carry out any analysis of portfolio mix at a more granular level than “retail/corporate”.
Also NB that the 5% weightings will be on seasoned mortgages, and a new lender’s book is by definition unseasoned. It’s not just a matter of developing historical data to validate the mortgage – an old loan is actually safer than a new one (in the limit, of course, a 25 year mortgage with six months left to run is genuinely almost risk free).
Fair point dsquared – although to be fair many of the new entrants are new entities buying seasoned portfolios (e.g. the RBS branches), so FSA is right to be generous to these entities.
I rather wonder what the distribution of RWs would look like if you applied the tools developed to understand US RMBS to the mortgage book. I don’t have a good sense, but presumably one would get a wide distribution of unexpected loss estimates of which the IRB ‘retail mortgage’ RW was an average?