New opportunities for gaming the system

The way minimum bank capital is calculated may be tweaked so comprehensively by the Basel Committee for Banking Supervision that the creation of a standardised third option is a distinct possibility. However, as the changes proposed also introduce higher levels of discretion into the calculation of some types of credit risk, they could undermine the main purpose of the Committee in shoring up bank capital, one high-profile financial services academic has argued.

Earlier this week the Australian Centre for Financial Studies released the latest instalment in its financial regulation discussion paper series, "Basel 3++: Buttons Belts and Braces", written by research director and FSI member, Kevin Davis.

Davis' paper considers the merits of two requests for comment on a proposed introduction of a capital floor for banks and changes to the standardised approach to credit risk, released by the Basel Committee for Banking Supervision last month.

For banks that use internal risk-based models to calculate their minimum capital requirements, the first proposal follows on from increases to minimum capital levels under Basel III and the implementation of a leverage ratio from the beginning of 2018.

Now unsure whether these measures are sufficient to ensure that large, more sophisticated banks are adequately capitalised, the BCBS is proposing a third measure based on a minimum capital requirement, determined from applying the standardised approach (which the BCBS now wants to revise anyway, as set out in the second request for comment).

The capital floor for a bank would then be the amount that is the greatest of the three measures applied. As. Davis points out, "This is reminiscent of a man uncertain of whether the buttons will hold his trousers up, and thus adds a belt and, just to be sure, attaches braces!"

Davis concludes, "... there is merit in reviewing whether the IRB approach of Basel 2 and 3 was an experiment which has, arguably, failed." His conclusion makes the logic of seeking to more closely align the standardised approach with the IRB approach questionable.

The proposal to revise the standardised approach to assessing credit risk, and therefore minimum capital requirements for less sophisticated banks that don't use the IRB approach, is aimed at reducing reliance on external credit ratings (a G20 goal stemming from the GFC); increasing risk sensitivity; reducing national discretion; strengthening the link between the standardised approach and the IRB approach; and enhancing comparability of capital requirements across banks.

The removal of credit ratings from the standardised approach is to be achieved by replacing look-up tables using credit ratings to determine the risk weighting of bank assets with look-up tables that use "risk drivers" instead.

The fact that the determination of some risk drivers will allow some national discretion appears to do nothing towards minimising national discretion in application of the standardised approach (another point seemingly lacking in logic).

Asset quality measures for banks, revenue measures (denominated in euros) for corporates and debt service coverage ratios for residential mortgage lending may all require levels of national discretion when being applied in less developed economies. These are the main asset types for which credit ratings will be removed as determinants of credit risk.

For exposures to sovereigns and their central banks, credit ratings will continue to be used to determine credit risk, and credit ratings will still be used for exposures to public sector entities, multi-lateral developments banks (supranationals) and collateralised exposures.

Presumably no alternative risk drivers have yet been identified.

For exposures to other banks, the risk drivers are the Common Equity Tier 1 (CET1) ratio and the Non-Performing Asset (NPA) ratio of the bank being assessed. If the bank has a CET1 ratio of 12 per cent or more and an NPA ratio of one per cent or less, the exposure can be weighted at 30 per cent.

However, if the bank has a CET1 ratio of 4.5 per cent or less, the exposure must be weighted at 300 per cent, regardless of what the NPA ratio is.

For corporate exposures, if a company has annual revenue of greater than €1 billion and leverage of up to three times, the exposure can be weighted at 60 per cent. Leverage is simply defined as total assets/total equity, as determined under national accounting standards.

If a company has annual revenue of no more than €5 million and leverage of five times or more, however, the exposure will be weighted at 130 per cent. And, if equity is negative, any exposure will be weighted at 300 per cent.

Of course, credit ratings are not used for assessing residential mortgage exposures but the BCBS is seeking to move away from applying a blanket 35 per cent risk weighting.

Weightings for residential mortgage exposures will be determined by the loan to valuation ratio for the mortgaged property and the debt service coverage ratio of the borrower (the latter being the proportion of mortgage repayments relative to the borrower's income).

A LTV ratio of less than 40 per cent and a DSCR of no more than 35 per cent will result in a weighting of 25 per cent, which better than the standard 35 per cent currently applied. Even with an LTV of greater than 100 per cent the weighting can be 80 per cent, so long as the DCSR is no more than 35 per cent.

If not, the weighting goes to 100 per cent.

Davis points out that both of these measures typically improve over the life of a mortgage. Therefore, there will be considerable incentives for borrowers to refinance in search of a better deal or for bankers to continually reassess their mortgage exposures. This is not necessarily a bad thing.

However, it is not clear that the proposed changes to the standardised approach go far towards achieving the stated aims of the BCBS. Rather it seems that the changes will just open up new opportunities for gaming the system.