The growth in derivatives trading activity in securities businesses has led securities firms to develop methods to analyse, control and report their trading risk in a consistent and reliable way. Firms have increasingly been turning to more sophisticated statistically based risk management methodologies using modern option and portfolio theory, and in particular to the use of risk modelling techniques, commonly referred to as value at risk modelling.
The increased use of these models, which seek to estimate with a specified degree of certainty the maximum loss which the firm would suffer in the face of possible adverse market movements. has led to calls for banking and securities regulators to recognise the output of these models for regulatory purposes. The Basle Committee on Banking Supervision (Basle Committee) has proposed the use of value at risk models for the purpose of calculating capital requirements for market risk, and the six major US derivatives dealers which make up the Derivatives Policy Group (DPG) in the USA are now providing value at risk information to the Securities and Exchange Commission (SEC) and the Commodity Futures Trading Commission (CFTC) as part of their arrangements for voluntary oversight. The output of value at risk models is also recognised within the framework of the European Union's Capital Adequacy Directive, which comes into force on I January 1996.
In the light of these developments, this paper outlines the issues for consideration by securities regulators when supervising firms which use such models, and considers in particular some of the implications of setting regulatory capital standards based on the output of internal models.
In Section 2, the paper explains how value at risk models are constructed. It explains the
difference between the calculation of value at risk, which is based on the assumption that events in the future will be similar to those in the past, and the results of stress tests which assess the potential loss a firm may face in certain extreme cases. The paper points out the role which models should play as an integral part of a firm's risk management procedures, and emphasises the important link which exists between the effective use of value at risk models and a firm's internal organisation and management controls. It also identifies the importance for regulators of analysing the firm's trading results in comparison with its value at risk calculations as a measure of the effectiveness with which the model is being used.
Section 3 considers the implications for securities regulators of recognising the output of value at risk models for the purpose tit calculating capital requirements for market risk. It notes the risk that firms could be incentives to seek out models which are most effective in reducing their capital requirements rather than those which are most effective for the purpose of better risk management. It also notes the problem which regulators face in determining how big a cushion of capital firms should be required to hold, over and above that implied by f.model, in order to cover unusually large adverse market events which are not captured by the modelling process. Other issues identified in this section include:
- the need for regulators to specify certain of the parameters for models which are to be used for the purpose of calculating regulatory capital;
- the implications for regulators of the fact that models provide more favourable results in larger portfolios; and
- the resourcing Implications for regulators of the need to approve models.
In Section 4, the paper notes that the Basle Committee invited IOSCO to comment on its April 1995 proposals on capital for market risk. It welcomes the work of the Basle Committee on the use of value at risk models and proposes that there should be further joint work between IOSCO and Basle on model testing. It also proposes that further consideration be given to the development of standards of best practice to be adopted by firms which wish to use value at risk models for regulatory purposes. However, in the light of the considerations outlined in the paper, the much greater significance of market risk capital requirements for securities firms, the lack of data regarding the reliability of such models in practice and the need to await the results of current initiatives such as the DPG in the USA, the paper notes that the Technical Committee does not feel that it would be appropriate at the present time to set a timescale for a move to the use of models for setting market risk requirements for securities firms.