In order to carry out a proper risk analysis, the client must give RiskSystem information about both their positions
and the manner in which they want the analysis to be carried out. There are two main sub-sections to Data Entry:
Position Data
and Preference Data.
The Position Data allows users to enter their positions on which the analysis will be carried out.
Clients of RiskSystem may hold their positions in their private account or on our servers.
The preference data is used to determine exactly the type of risk analysis that the user wishes to carry out.
There are a considerable number of options however all users are assigned default values that are consistent
with current market practice. These preferences will only be stored if the client registers with RiskSystem.
At RiskSystem, there are four main methodologies that are used to determine the risk of a clients portfolio.
These are (1)
Gross Exposure, a static analysis based purely on
position value.
(2) Stress Testing, a static analysis based on position
value and arbitrary market moves.
(3) Value at Risk (VaR), a dynamic analysis based on position
exposure and market dynamics. This is the most common methodology used for risk analysis in finance and carries the imprimatur of the Bank of
International Standards, whom are responsible for the Basle Requirements.
(4) BDaO Analysis, an combination of VaR and Historical
stress testing designed to determine the impact of correlation assumption on worst case portfolio dynamics.
At the core of risk analysis is the assumption that it is possible to forecast the extreme dynamics of a portfolio,
based on the historical dynamics of the underlying constituents of the portfolio. The stability of the portfolio construction is tested in two separate methods.
Portfolio Construction is analysed under a number of methodologies such as Mean Variance and Ledoit Shrinkage.
Extreme dependence on the method of construction lead to significant sample errors which degrade the value of the risk analysis.
Bregman Risk Measures, are dynamic analyses based on position exposure,
market dynamics and portfolio internal dynamics. They can give statistically significant signals indicating when a portfolio is moving away from its
original form, before this is observed in the actual performance of the portfolio.