This series of blog posts explains about Quartet FS’s value at risk software and how to get the most out of our online VaR demo. In this post, we explain different types of value at risk analysis and the cube structure in the demo.

• The value at risk software demo illustrates how three different kinds of Value at Risk can be analyzed:
• Historic Value at Risk looking back over 500 days (typical for 2 years history). On this analysis the confidence level is configurable on a per-user basis: typical values are 99% or 95%.
• Stress Value at Risk where the results of 250 different stress analyses are to be aggregated and analyzed by the Value at Risk software
• Scenario Value at Risk where the results of 250 different scenarios are aggregated and analysed to locate the worst scenario (equivalent to the 100th percentile)
• This value at risk software project demonstrates the calculation and aggregation of various value-at-risk measures for a multi asset class portfolio across many dimensions and levels. It computes value at risk for historical profit and loss simulations based on historical data, stress test and specific scenarios produced by a separate risk engine.
• Data is provided as files containing trade and positions attributes (e.g. book, counterparty, trade type) and vectors of 1000 profit and loss simulations per risk per trade/position (split across 500 historical, 250 stress and 250 scenarios p&l values) , plus some reference tables used to enrich trade attributes.
• This value at risk software demonstration highlights the capability of ActivePivot to:
• Compute and aggregate value at risk on a large portfolio with many dimensions,
• Slice and dice value at risk across many trade attributes,
• Visualize details of calculations per trade and original scenario
• Compute complex additional measures (quality level, works cases, marginal var) to better manage results.

## Cube Structure

Most dimensions are self explanatory. However a few of them are worth explaining further: AsOfDate represents the business day time line. It is possible to hold and compare several days’ worth of data.

RiskType breaks down value at risk across the four main asset classes: IR (interest rate), FX (foreign exchange), EQ (equity) and CS (Credit). A second levels shows detail to a finer type of risk, such as credit spread versus exposure to default for Credit assets.

Validity is used to rapidly identify where calculation problems have been encountered, such as missing trade data by grouping cells into two categories: CORRECT and MISSING ATTRIBUTES. By using other dimensions as cross join or doing a drill through it is very easy to find out which trades are causing problems.

Var detail projects the details of the aggregated P&L vectors on rows or columns. Beware that there will be a lot of rows or columns returned by the server to the front end display (1000). In the next post we will explain about the different measures used in the Value at Risk demo.