VaR-analysis Module
VaR-analysis Module is designed for VaR risk assessment in portfolio management. It allows evaluating losses which would not exceed a specified probability (confidence level) after market risk occurs. The evaluation in money terms and percentage may be performed for a particular asset and portfolio over a chosen time period.
The module can be applied to the following assets:
- stocks,
- bonds,
- futures,
- options.
Calculation methods
Parametric:
- Delta-normal
The basic modelling method uses a hypothesis of normal yield distribution.
For more accurate modelling of ‘heavy tails’ in yield distribution the following method modifications are used: - Delta-gamma normal — CORNISH-FISHER (the 3rd moment)
- Delta-gamma normal — CORNISH-FISHER (the 3rd and 4th moments)
VaR is calculated separately for each asset type.
When evaluating volatility with parametric methods the following approaches are applied to calculation of covariance matrixes:
- SMA methods with constant covariance:
- average revenue approach,
- zero average revenue approach.
- EWMA method with more weight given to recent changes of risk factors:
- zero average revenue approach.
Historical:
- Basic
- Bootstrap
The Module calculates the current portfolio status taking into account recorded historical data. Calculation results are provided in the form of reports.
The historical methods allow making general VaR portfolio evaluation with details by asset types.
The Module calculates and reports a few additional parameters for each asset type:
- for stocks — instrument share in portfolio, VaR in percentage and cash terms, beta, VaR sensitivity to change of instrument share (marginal VaR), contribution of instrument to portfolio VaR (component VaR);
- for options (for every strike) — option greeks, position greeks, position delta and gamma in money terms, VaR and other position risk indicators;
- for bonds — modified duration, «convexity», yield to maturity, standard yield deviation, VaR and other indicators.