To aggregate market risks across diverse asset classes, organizations use Value at Risk (VaR) models. VaR provides an estimate of the maximum potential loss over a specific timeframe within a defined confidence level. Risk teams use different methodologies to calculate this metric:
The Parametric (Variance-Covariance) VaR Model
The parametric approach assumes that asset returns follow a standard normal distribution. It uses historical price data to calculate the mean and standard deviation for each asset, then applies correlation metrics to estimate portfolio risk.
- Strengths: Highly efficient to calculate and requires minimal computing power.
- Weaknesses: Fails to predict non-normal market events and can underestimate the frequency of extreme tail-risk losses.
The Historical Simulation VaR Model
The historical simulation approach does not assume a normal distribution. Instead, it applies actual price movements from a past historical window (e.g., the preceding 250 trading days) directly to the current portfolio to calculate potential returns.
[Current Portfolio Assets] ---> Apply Past 250 Days of Real Price Changes ---> Sort Outcomes Low to High ---> Extract Target Percentile Value
This model naturally captures real-world market distributions and tail-risk adjustments, making it a reliable tool for risk analysis during volatile market cycles.
Â