8. To assist in this investigation and to identify potential supervisory concerns, a Committee task force studied the market risk models and management practices of a number of banks in the major financial centres. In particular, it carried out some preliminary testing in the second half of 1994, in order to help determine which model parameters should be specified or constrained. One objective of the test was to check whether the banks' internal measurement systems would produce, in the Committee's view, reasonable value-at-risk estimates relative to the size of the portfolio. Another was to establish how great would be the dispersion between different models' measures of value-at-risk when relatively few parameters were specified. The results of these tests have guided the choice of quantitative parameters now being proposed by the Committee, although these are not definitive and may be subject to changes in the light of further work. Through this testing, the task force has recognised the scale of the human and technological resources that many banks are devoting to the development of their market risk models, and the Committee has gained a heightened appreciation for the rapid pace of change within the industry which is prompting banks to make these major investments in resources.

9. As a result of the investigation conducted on internal risk measurement models, the Committee has decided to envisage, subject to a number of carefully defined criteria as set out in Part B of the proposed Supplement, the use of internal models to measure market risks for supervisory purposes. The issues raised by this decision are examined in a separate paper "An internal model-based approach to market risk capital requirements".

10. The framework for the use of internal models will contain both qualitative and quantitative standards. The general elements can be summarised as follows. For the qualitative standards, it is required that there should be an independent risk control unit with active involvement of senior management in the process, the model must be closely integrated into day-to-day risk management and a routine and rigorous programme of stress testing should be in place. Banks must have a routine for ensuring compliance and an independent review of both risk management and risk measurement should be carried out at regular intervals. In addition, procedures are prescribed for internal and external validation of the risk measurement process.

11. For the quantitative standards, the "value-at-risk" should be computed daily, using a 99th percentile, one-tailed confidence interval and a minimum holding period of ten trading days. The historical observation period will be subject to a minimum length of one year, but the Committee is also investigating the possibility of a dual observation period. Banks will have discretion to recognise empirical correlations within broad risk categories, but value-at-risk across these risk categories is to be aggregated on a simple sum basis. Models must also accurately capture the unique risks associated with options. The capital charge will be the higher of.

• the previous day's "value-at-risk";
• an average of the daily "value-at-risk" on each of the preceding sixty business days, multiplied by a multiplication factor assessed by each national supervisor in accordance with the criteria set out in paragraph 12 below.

12. The multiplication factor- will be set by individual supervisors on the basis of their assessment of the quality of each bank's risk management system, subject to an absolute minimum of 3. However, in considering the use of models for supervisory purposes, the Committee has been very conscious of the need to provide banks with both the flexibility and the incentives to upgrade their internal models as financial markets and technology evolve. The Committee has therefore agreed that banks should be required to add to this multiplication factor a "plus" directly related to the ex-post performance of the model, thereby introducing a built-in positive incentive to keep high the predictive quality of the model (e.g. it could be derived from the outcome of so-called "back-testing" and be zero when such results are satisfactory). More work will be done during, and on the basis of, the consultation process to check further the feasibility of the "plus" and to arrive at a more precise definition of it.

13. Since the use of proprietary in-house models to measure market risk for supervisory purposes represents a significant innovation in supervisory methods, implementation of the approach will of necessity be to some extent evolutionary. The Committee accordingly reserves the right to modify the specifications required for banks using models as more experience is gained. During the consultation period, it intends to conduct a second test exercise using the parameters now proposed and will examine those results in reviewing comments from the industry. The Committee will seek to ensure that the dispersion of results across institutions for a given set of positions falls within a reasonable range, and it will work with the industry to achieve this goal. Moreover, in order to gain additional information and comfort with the results produced by internal models, supervisors reserve the right to require banks wishing to use internal models to perform testing exercises and to provide any other information necessary to check the validity of banks' models. All banks that wish to use models should therefore have the capability to evaluate a test portfolio.