Microstructure & Parametric VaR: Institutional Risk Management

Beyond Simple Stop-Losses

Retail investors manage risk with arbitrary stop-losses. Institutional quantitative systems manage risk using covariance matrices and Parametric Value at Risk (VaR).

The 99% Parametric VaR Model

When our engine proposes a portfolio of 5 to 10 equities, it downloads the trailing 60-day price history for every asset. It then calculates the covariance matrix to understand exactly how these assets move together.

• Correlation Caps (r > 0.75): If two stocks are highly correlated (e.g., AAPL and MSFT), the engine automatically drops the lower-ranked asset. We refuse to pay for false diversification.
• 99% VaR Calculation: The engine calculates the maximum expected loss over a 10-day period with 99% statistical confidence. If the portfolio VaR exceeds our internal threshold, the position sizing is mathematically scaled down before publication.

By enforcing microstructure and VaR constraints, the portfolio is mathematically insulated against systemic shocks.