Newsletter February 2005
New worldwide study to test Basel II impact
Recalibration decisions brought forward at US request. Race to include results of trading book review in coordinated QIS5 plan By Melvyn Westlake
Regulators on the Basel Committee are planning another full-scale, internationally coordinated test of the impact that the new risk-based capital rules will have on the total level of capital in the world's banking system. This plan for a fifth quantitative impact study - or QIS5 - is a significant departure from previous intentions. It brings forward by several months the timetable for making any crucial adjustments needed to achieve the desired level of regulatory capital in the system or held against particular banking businesses.
However, the QIS5 plan, which was first decided at the mid-December meeting of the Basel Committee, architect of the new capital regime - Basel II - has created a good deal of confusion among bankers. Many of the details were left open. As Global Risk Regulator went to press, a meeting of the Capital Task Force, an executive sub-committee, was taking place at the Bank of England to hammer out specific aspects of the plan.
The purpose of the impact study is to take a snapshot of banks' balance sheets in order to assess whether there should be a recalibration of the various risk weights and curves that determine how much regulatory capital banks need to hold to underpin their solvency. Regulators have insisted that the introduction of the new risk-sensitive Basel rules should not lead to any overall reduction in the capital of the banking system.
Previously, it had been intended to undertake the recalibration exercise during 2006 when banks will be running their capital models and risk systems under the Basel II rules in parallel with their existing Basel I procedures. It had been assumed that a balance sheet snapshot based on this parallel running would be taken around the middle of that year. This programme has now been changed, largely at the request of the US. Washington bank regulators want to include fairly definitive numbers, taking account of any recalibration, in their own rule that puts Basel II into effect in America. A Notice of Proposed Rule-making (NPR) is to be issued in June. This will be followed by a period of consultation, and the promulgation of the final US rule around the middle of next year.
September spreadsheets
The precise dates on which banks around the world will now undertake the QIS5 balance sheet snapshot of their capital levels is being left to national regulators, although September 30, and December 31 are among the most likely. Despite the coordinated nature of the exercise, it is not necessary for all participating banks to do it on the same day.
However, it was decided at the Bank of England meeting of the Capital Task Force on February 17-18 that the Basel Committee secretariat will deliver the necessary spreadsheets and workbooks to national regulators in September. These will then have to be submitted to the Basel committee, duly completed, by March. Allowing time for examination and reconciliation of the completed spreadsheets, final decisions about any recalibration are expected to be taken during the second quarter of 2006.
Other aspects to be resolved included how extensive the exercise will be. It will include key banks in all of the 13 developed countries that are members of the Basel Committee. But some regulators suggest that it should also include as many other countries as possible. The last coordinated impact study, QIS3, conducted at the end of 2002, involved over 360 banks in more than 40 countries.
"There are no restrictions on which countries can participate," says Gerhard Hofmann, head of the banking supervision department, at the Bundesbank in Frankfurt, and chairman of the Basel Committee's OCQIS (Overall Capital Quantitative Impact Study) sub-committee organising the exercise. He is planning one or two seminars this summer to provide guidance to countries that want to take part.
However, the various individual tests and studies carried out by some countries have confused the picture. For example, both the US and Germany have just undertaken individual QIS4 exercises, and it is not clear whether they will now carry out full-scale QIS5s, or something more limited and partial. Britain had somewhat reluctantly announced that it, too, would undertake a QIS4. Although the Financial Services Authority, Britain's regulator, may encourage banks to go ahead with this exercise, it will put its main effort into QIS5.
As the Basel II project manager at one leading British bank says: "It confuses the hell out of me." Some supervisors have also been obliged to change their plans. That has not made everybody happy, notes one supervisor. But there is really no other choice if calibration is to be considered in a coordinated way by all members of the Basel Committee, including the Americans, he says. "Countries will have to make decisions pretty quickly after the Bank of England meeting."
There are several crucial elements that could make the QIS5 exercise particularly important in any consideration of recalibration. One is the inclusion of results of the trading book review, although it now appears that only preliminary results will actually be available in time for the impact study. By bringing the recalibration process forward several months, regulators are, at the same time, having to squeeze the timetable for the trading book review. This review was set in motion last year when it became clear that it would not be possible to resolve the capital treatment of certain trading book activities in time for inclusion in the Basel II accord text, published last June.
The hope is that the trading book review can be completed in March or April this year. Completion of the review will be followed by a public consultation process. Several further weeks will then be needed for the comments to be considered. So, the final text is unlikely to be ready by September.
However, inclusion of the trading book review, even in a preliminary form, is viewed as hugely important to some of the world's largest banks. One major banking group, which is expecting only a very minimal reduction in capital on the basis of last June's Basel II text, reckons that inclusion of the trading book review results will produce a significantly bigger capital reduction.
Other factors
This is not the only factor that could make the outcome of QIS5 significantly different to that of QIS3. A second factor will be the inclusion of a definition for so-called stressed loss-given-default calculations. In addition, QIS5 will be the first coordinated impact exercise since the decision to base capital models on unexpected losses only, rather than on both expected and unexpected losses.
There is also a fourth factor that is potentially more important than any of these other three factors, says one European regulator. The QIS3 exercise involved a lot of guesswork. Quality control of the data often left a lot to be desired. Banks will be running QIS5 quite close to the way they will be running the actual systems when they go live, whereas it was not real systems producing the QIS3 numbers, says the regulator.
Different outcome possible
"There is quite a potential for things to turn out differently using real QIS5 systems, compared with the `pretend' QIS3 systems. And, these four new factors, taken together, could potentially move the numbers [for risk-based capital] quite significantly," he concludes.
It doesn't follow that there will necessarily have to be a recalibration, regulators caution, nor any resort to the scaling factor. The Basel Committee has introduced the concept of a scaling factor so that it can, if necessary, make an across the board, linear adjustment to the level of capital required under Basel II, preventing an excessive and unpredicted rise or fall. After QIS3, the committee provided an indicative scaling factor of 106% (ie. a 6% overall increase).
The scaling factor is the more likely adjustment mechanism because it would not run the risk of altering the balance or relationship between specific bank portfolios, such as retail lending or corporate lending, says Gerhard Hofmann. But if QIS5 throws up anything unusual or strange, the Basel Committee may have to consider specific portfolio recalibrations. The committee would then have to weigh the trade-offs between making such specific portfolio recalibrations and making alterations that risk unbalancing the entire accord, Hofmann says.
"An unexpected rise or fall in capital in a business line may well be taken as evidence that the new Basel accord is not delivering an appropriate risk-sensitive level of capital in that particular area," says another European regulator.
Slippery slope
"And that could lead to a call to redesign that particular area of the accord. Then you quickly shift out of recalibration into redesign. That is a slippery slope," he says. On the other hand, a significant unexplained change in a particular portfolio could not just be ignored, the regulator adds.
Meanwhile, work on the trading book review is said by observers to be "proceeding reasonably well." Workable compromises that satisfy both regulators and industry representatives appear to be emerging from two of the three working groups. The working group dealing with counterparty credit risk has made particular progress. Proposals also look very likely to be produced by a second working group, dealing with double default - where debt has effectively got two obligors and may be viewed as having reduced risk - and securities of short-term maturity.
There is rather less progress in a third working group, dealing with specific market risks arising from concentrations of illiquid assets, such as bank holdings of structured financial products. This matter is dealt with fairly crudely under the 1996 market risk amendment to Basel I. But some regulators want to see a more sophisticated modelling approach adopted by banks.
At the outset it was intended that the three working groups should complete their work at the same time, because the proposals from the first two are likely to result in capital reductions for banks and investment banks, while proposals from the third working group is likely to lead to an increase in the capital requirement. This should lead to a balanced overall package.
However, strong differences have emerged between regulators. Some regulators still insist that the results of all three working groups should be delivered together as quickly as possible, hopefully in an April consultation paper; some others think the work of the third working group should be continued separately after the first two have delivered their proposals; and yet other regulators are believed to have little enthusiasm for the third working group, anyway.
One observer predicts that an April consultative paper will contain proposals from all three working groups. But he doubts that all the proposals from the third working group will be sufficiently clear, detailed and precise to be incorporated into Basel II text or used as an input into the QIS5 exercise, without more work and yet a further round of consultations. The situation may be clearer after a week-long meeting of the three working groups in Madrid, scheduled to begin on February 28.
