Basle Survey Signals Focus Of Discussion Paper

BASLE, SWITZERLAND -- The focus of the operational risk discussion document planned by global banking regulators is signalled in a survey seeking information on banks’ losses from such hazards as fraud, computer system failure and trade settlement errors.

The survey, the so-called second quantitative impact study (QIS 2), was issued by regulators to major banks in early May as Operational Risk went to press.

Its primary purpose is to collect in confidence "event by event" -- or granular -- data on losses suffered by banks over the past three years, so regulators can decide on the form of the internal measurement approach for calculating an op risk capital charge.

The Basle Committee on Banking Supervision, which in effect regulates international banking, wants the data returned by August 1. It acknowledges the survey is lengthy and detailed and that the deadline is tight.

But the committee, which comprises banking supervisors from the Group of 10 leading economies, says the data requested "is of fundamental importance to the development of the operational risk charge." The charge is proposed under the proposed Basle II capital adequacy accord due to come into force in 2004.

The data is needed to consolidate banking industry-wide operational loss data on a consistent and comparable basis. At present, there is no standard loss classification and measurement system for op risk.

Meanwhile, the committee plans to issue its special discussion paper in July. Regulators say the document will outline the ways the loss data being collected by the QIS 2 survey could be used to calculate an op risk charge using banks’ own internal measurements.

The internal measurement approach, which will rely on banks’ own loss experiences, is the most complex of the three approaches to calculating an op risk capital charge under the Basle II accord.

The July discussion paper will try to bring together the various ways of thinking about the internal measurement approach, say regulators. They add that this could well result in options being discarded.

The Basle II accord, put forward by the Basle Committee in January for comment by May 31, will require large international banks to set aside capital for the first time against the risk of losses from operational hazards.

Under the three-way choice of methods for calculating op risk capital charges offered under Basle II, the more complex the method a bank uses, the lower will be its capital charge.

But bankers say very few of even the largest banks are yet in a position to use the internal measurement approach, the most complex of the three methods.

The European Commission proposes bringing in similar rules, closely modelled on Basle II, for all banks and investment firms in the 15-nation European Union by the same date.

Meanwhile, bankers still have to meet the June 1 deadline for the return of information sought by the Basle regulators in the first tranche of their quantitative impact study -- QIS 1 -- which was issued in mid-April.

The primary purpose of QIS 1 is to gather data to assess the overall impact of the Basle II proposals on a wide range of banks in the G-10 and beyond.

But QIS 1 also seeks information that regulators can use to estimate the impact of the simplest of the three approaches to calculating the op risk charge -- the basic indicator approach.

The basic indicator approach is likely to calculate the charge simply by multiplying a bank’s gross revenue by a factor, alpha.

QIS 1 also seeks to help with the further development of the standardised approach, the second most complex method, as well as the internal measurement approach.

The standardised approach divides a bank into business lines and applies a factor, beta, in calculating the charge.

The internal measurement approach also divides a bank into business lines, such as retail banking, asset management and payment and settlement. And, as with the standardised approach, each business line is given a standard risk indicator -- for example, annual average assets, total funds under management and annual settlement throughput.

But the internal measurement approach also defines a set of operational loss types -- write-downs, loss of recourse, legal liability, loss of assets, for example -- and applies them across business lines. Regulators will supply a factor, gamma, for each business line/risk type combination that will be used to calculate the op risk charge. The loss data is needed to calibrate the gammas.

QIS 2 and its accompanying spreadsheet set out a format for submitting loss data. The format is based on eight standard business lines, which are split into two levels. The first level contains seven loss event categories, and in turn is sub-divided into a second level of 21 loss events.

To ease the reporting burden, the regulators propose that the data requested in QIS 2 should be submitted in electronic format, using an Access database template. The template will be provided in June, but QIS 2 is being circulated before then, so that banks can plan their data management and collection exercises.

The regulators says it is clear that even those banks with well established reporting systems will need to map their existing data to present it in the format requested. They add that banks without loss databases are unlikely to be able to collect and present data within the time set.

But the regulators are urging all banks to review QIS 2 and think about how they might establish mechanisms to collect such data in the future.

The Basle Committee will conduct further data exercises on op risk next year, based on the QIS 2 format. It hopes in this way to lay the foundations for improved data reporting and capture by banks, and hence more accurate calibration of the op risk charge.

Regulators say it is important that banks complete the survey using the definitions provided. Where banks’ definitions do not correspond to those in the survey, they should map them to those provided in QIS 2.

Banks sending in results that do not follow the QIS 2 definitions and mappings will be discarded from the sample. "If you can’t provide the loss data, then don’t fill in the form," said one regulator.

Banks submitting loss data must also provide related information on exposure indicators and scaling -- the adjustment of loss size and rates to bank size -- so that the information can be aggregated with that provided by other banks.