var-backtesting

Measure market risk by CAViaR model

A regime-switching Monte Carlo engine for modeling systemic risk and contagion in equity markets.

Quantitative market risk model monitoring framework implementing 99% one-day VaR using Historical, Parametric, EWMA, and Monte Carlo methods. Includes rolling out-of-sample forecasting, Kupiec and Christoffersen backtesting, exception monitoring, stress testing, volatility and drawdown analysis, and model validation using Python and Excel.

VaR backtests (Kupiec, Christoffersen, Basel traffic light) with exact finite-sample p-values