1   Risk Control Assets
          The user can select the assets to be used in the risk control functionality.
2   Risk Control Assets from Portfolio
           The user can select assets from the portfolio to be used in the risk control functionality.
1   Hedge Strategy
          This allows the user to determine the impact of an optimal hedging strategy on the risk of their portfolio A variety of user specified instruments may be selected. Optimal in this context means that which reduces the risk most.
2   Risk Targeting
           The user is able to select a target risk level and then using a variety of user selected instruments, the most efficient target portfolio will be determined. The risk of the user portfolio and the target portfolio will be as close as is possible to get to the target risk level. In this case most efficient means that if the target level is attainable, the target portfolio selected will be that portfolio which has the lowest risk. In the situation where the target is not attainable, it will be that portfolio that, in combination with the user portfolio, most closely approaches the target risk level.

Gross Exposure: Summation

Total Gross Exposure may be calculated in two ways. Firstly, by summing the absolute Gross Exposure of the individual positions. Secondly by summing the actual gross exposures of the individual positions, thereby potentially reducing the TGE due to the offsetting inpact of long and short positions.

Gross Exposure: Percentage

The Gross Exposure (absolute or actual calculation) may be used as a risk metric as a function of assets under management (leverage). However in the event that it is being used as an estimate of potential loss risk metric, a percentage of the Gross Exposure may be a more appropriate calculation

Stress Test: Summation

Total Gross Exposure may be calculated in two ways. Firstly, by summing the absolute Gross Exposure of the individual positions. Secondly by summing the actual gross exposures of the individual positions, thereby potentially reducing the TGE due to the offsetting inpact of long and short positions.

Stress Test: Equity

This is the user specified estimate of an extreme move in the Equity Asset Class. It will be applied to both indices and stocks. If a client would like more options please contact us directly.

Stress Test: Commodity

This is the user specified estimate of an extreme move in the Commodity Asset Class. It will be applied to both indices and stocks. If a client would like more options please contact us directly.

Stress Test: Forex

This is the user specified estimate of an extreme move in the Foreign Exchange Asset Class. It will be applied to both indices and stocks. If a client would like more options please contact us directly.

Stress Test: Fixed Income

This is the user specified estimate of an extreme move in the Fixed Income Asset Class. It will be applied to both indices and stocks. If a client would like more options please contact us directly.

VaR: Methodology

VaR may be calculated using historical data, where actual historical price movements are used to determine implied profit/loss of a portfolio over a range of dates. Alternatively, a variance / covariance matrix (determined using historical data or from a template) and a randon number generator may be used to simulate historical prices. Currently at RiskSystem only Historical Var is provided to clients. If you require Simulated VaR please contact us directly.

VaR: Distribution

the VaR of a time series is calculated from a probability distribution of historical prices. The distribution may be modeled using a function (parametric) or directly from the data (non-parametric). In addition it is possible to consider only extreme losses (one-sided) as opposed to both extreme profits and losses (double-sided) to forecast potential future price movements.

VaR: InSample Window

This is amount of historical data that will be used in the calculation. The longer the time window, the greater the potential range in price movements. However, such movements may lack relevenace to current market conditions. For reference, the Bank of International Standards recommends a 1 year unweighted (Dirichlet) time series.

VaR: Window Function

The window function weights the time series data accordingto how they fall in the series. It has the effect of empasising or de-emphasising the impact of recent data, and data close to the start of the time series. It has particular relevenace when comparing the evoluition of a position VaR over time, where the impact of new data, and data points leaving the time window may result in a very erratic measure of risk.

VaR: Confidence Interval

This is a measure for how often one would expect a profit or loss, greater than or equal to the risk measure, to occur. Given the fat-tailed (leptokurtic) nature of financial time series, and the generally gloomy nature of risk analysis, it would be wise to assume that bad things more often than forecast...

VaR: Valuation

This allows the user to determine whether of not to include second order effects such as quantos etc in the detewrmination of risk. In the event of the client having options (puts, calls, non-guaranteed stops etc) in their portfolio, the default option will be delta-gamma.

VaR: Distribution

In the event of an asset having a current price close to zero, historical price moves may imply an asset price less than zero (Normal). This may be avoided by truncating all moves that imply a negative asset price to a move that imples a zero asset price (Truncated Normal). An alternative is to createa VaR of returns and using that to generate the price differences from the current price.

VaR: OutSample Window

This determines the number of days over which the forecast is made.

VBDaO: Type

The aggregation in the BDaO process may be done over the asset price changes directly in a given risk sector, or over the VaR of the asset prices in a given risk sector. The principa difference is in the stability of the evolution of the BDaO metric over time.

VBDaO: Type

The BDaO process generally looks at a much longer time series than is used in VaR so that a proper estimation of a "Bad Day" may be determined. If the VaR type is selected the InSample Window should be at leat 3-5 time s the length of the VaR insample window.