The quality of the market risks disclosure policy for the firm's trading book should be evaluated with respect to a standard presentation format which is shown in the tables below and based on the following set of quantitative and qualitative information:

a) Value-at-risk information

In Table 1, Panel A, the VAR statistics for the entire portfolio are provided and include the 10-day horizon VAR of the portfolio at the 99 % confidence interval by quarter (and as of the reporting date). The statistics also include the average, minimum and maximum VAR, over the period. Since they are required by the BIS, a 10-day horizon and a 99% confidence interval are likely to be the standard for comparison. In light of the statistical limitations of the scaling of the VAR daily figures for longer horizons, we Table 1 should be provided for daily VAR figures as well. The aggregate portfolio VAR is then broken down by major risk category for detailed comparisons.

Table 1
Trading Portfolio VAR 10-day versus 1-day VAR for a 99% confidence interval
Panel A - 10-day VAR
As of Dec. 31,	Quarter 1	Quarter 2	Quarter 3	Quarter 4	Average for year	Minimum	Maximum
Total Portfolio
Interest Rates
Currency
Equity
Commodity
Panel B - 1-day VAR
As of Dec. 31,	Quarter 1	Quarter 2	Quarter 3	Quarter 4	Average for year	Minimum	Maximum
Total Portfolio
Interest Rates
Currency
Equity
Commodity

The quantitative information set is supplemented in particular with:

• the description of the main assumptions of the various risk models (Monte-Carlo or historical simulation model, variance-covariance model, etc.) used by market risk factor type (equity, foreign exchange, interest rate and commodity),
• the main assumptions used to compute the volatilities and correlations (time horizon, weighting procedure, etc.) discussed for each major risk factor type,
• the method of aggregation across risk factors (simple aggregation, correlation based, etc.),
• the treatment of non-linear payoffs and, in particular, derivative securities by risk factors (as well as, for cross options, the handling of correlated risk factor exposures),
• the scaling method to compute the n-day VAR. Due to the limitations of scaling, a second panel B in Table 1 should present the daily VAR figures,
• the reference period used to compute the VAR estimates,
• the internal process for the aggregate and risk-factor-specific VAR models validation and verification.

b) Backtesting information

The annual report should comment, in particular, on the description of the backtesting method used (static versus dynamic method and based on the BIS exceptions rule or not). It should present a VAR frequency distribution chart or a figure comparing the actual daily P&L with the one-day VAR estimate and discussing the number and magnitude of exceptions. It should present the magnitude and frequency of significant changes in the various trading books composition and their impact on the backtesting procedure. Finally, the annual report should compare the frequency of the daily (and 10-day) VAR distribution with the economic capital allocated

c) Stress testing

Currently, there is little meaningful information disclosed on stress testing methodologies. The recent financial crises in Asia, Russia and Brazil reinforce the need to evaluate whether the firm monitors the overall exposure of the trading books to major changes in macro-economic variables. Examples include sharp increases in inflation and interest rates, twists in the shape of the yield curve, devaluations and sharp drops in one or several equity markets and related impact on market correlations. The most important message stemming from stress testing disclosure should be to identify whether a firm seriously assesses extreme events and the limits of its VAR models, in particular with respect to the composition of its trading books.

The sensitivity of the aggregate and major constituents of the P&L to scenario analysis based stress testing procedures should be disclosed. The firm's method should be discussed, including the motivation for the following choices: the stress factors, the historical period used to estimate the likelihood of market factor jumps, the impact of market disruptions on correlations across risks and the method to compute the impact of large moves on derivatives positions.