Global regulators have today agreed a new deal to impose a 7% Tier 1 capital hurdle for banks to reach by 2019.
This is aimed at increasing bank stability and providing a larger cushion against losses.
The issue that seems to have gone under the radar is 7% of what?
Raising the Tier 1 percentage to 7% is meaningless without a revisiting of Internal Risk-Based methodologies and the computations that go into the definition of banks’ total risk, of which 7% must be held.
Banks have increasingly, under Basel II, been allowed to self-certify the risks in their books of business, especially when there was real estate collateral.
The new measure certainly isn’t 7% of outstanding loans; it is 7% of the “risk-adjusted” face value of every piece of business the bank does that is considered to involve risk.
The risk-adjustment is an assessment of the size of the risk, meaning the likelihood of the risk materialising, multiplied by the impact if it does, and this goes across all the types of risk – credit risk, market risk, operational risk…
The likelihood of materialisation is more or less a function of the counterparty with whom the piece of business was transacted, such that a derivative done with, for example, the Kingdom of Sweden is far less likely to have to be replaced during its term than one contracted, for example, with a corporation.
The impact is the amount the bank could lose if the counterparty defaults, and is a function of the maturity date, the type of transaction, the security etc.., so a 15-year loan with a balloon repayment and no security fares worse than a 5-year loan with equal semi-annual repayments and security.
Each piece of business has its risk-weighted value: that percentage of face value of which the bank will have to have 7% Tier 1 capital, as opposed to 4% currently.
The problem is that these Internal Risk-Based methodologies for defining the risk-adjustment to face value have proven to be flawed, not least because they have been based on external data (e.g. on consumer mortgage defaults) that did not go back as far as the last recession.
An example of risk-adjustment under Internal Risk-Based methodologies would be to allocate a 10% risk-adjustment to a consumer mortgage where the loan was 85% of the property value. In other words a loan of 100 would have a risk-adjusted face value of 10, and a capital quotient 7% of 10 = 0.7: below 1% of face value.
The rationale would be that the bank would only lose money if (a) the borrower defaults and (b) the bank repossesses and re-sells and (c) property prices have fallen more than 25% since the time the loan was made.
Between 1993 and 2008 property prices only went up; on that basis one could even avoid worrying about the borrower going into default, hence self-certified mortgages
But we have seen loan losses much higher than this in Sub-Prime and sporadically where banks have repossessed, rather than capitalising interest, deferring payment and other measures to keep the loan “current” in accounting terms.
For example, The Prince of Wales purchased Dumfries House in Ayrshire in 2007 for £27m with a loan of £20m (75%) and with a view to redeveloping the estate. Now the value has dropped to an estimated £7m – but the loan has not defaulted. The Prince is reputed to be a good credit risk, so a foreclosure by the lender appears unlikely. But if the lender did repossess, their write-off would be £20m or 100% of the loan.
The bank’s internal capital adequacy computations remain so flawed, particularly on real estate lending, that raising the Tier 1 quotient to 7% does not give assurance that institutions will be stable in an absolute, rather than a relative, sense
PDF of Business Money September 2010