Introduction: The Mathematics of Tail Risk and Solvency

In institutional finance, capital preservation under catastrophic shocks is paramount. Following crises such as 2008, global regulations (Basel III/IV) mandated rigorous quantitative risk management.

Modern risk management goes beyond variance—it models non-normal distributions, tail risk, liquidity horizons, and multi-factor stress scenarios. This lesson covers VaR, Expected Shortfall, Extreme Value Theory, and stress testing frameworks.

 

Part 1: Value at Risk (VaR) Foundations

1. Formal Definition

VaR at confidence level 1−α over horizon T is the threshold loss L such that:

P(ΔP≤−VaRα)=α

Example: A 1-day 99% VaR of $10M means a 1% chance of losing more than $10M in one day.

2. Methodologies

  • Parametric (Variance-Covariance) VaR Assumes Gaussian returns:

VaRα=−(μ+zασ)×W0

Fast but underestimates fat tails.

  • Historical Simulation VaR Revalues portfolio using past returns. VaR = empirical α-percentile. Captures fat tails but blind to novel shocks.

  • Monte Carlo Simulation VaR Generates thousands of paths via SDEs (GBM, jump-diffusion). Flexible but computationally heavy.

 

Part 2: Beyond VaR – Expected Shortfall and Tail Risk

1. Expected Shortfall (ES / CVaR)

Addresses VaR’s flaws by measuring average loss beyond the threshold:

ESα=−1α∫0αVaRp dp

ES is coherent (subadditive, monotone, homogeneous, translationally invariant). Basel standards prefer ES.

2. Extreme Value Theory (EVT)

Models heavy tails directly using Generalized Pareto Distribution (GPD). EVT estimates probabilities of extreme losses, improving precision during crises.

 

Part 3: Institutional Stress Testing

1. Historical Scenario Stress Testing

Replay past crises (1987 crash, 2008 Lehman collapse, 2020 COVID freeze) against current portfolios.

2. Hypothetical Macroprudential Stress Testing

Synthetic forward-looking shocks mandated by regulators (CCAR, DFAST, EBA):

  • Rates +300 bps

  • Equities −40%

  • Credit spreads +500 bps

  • Liquidity freeze

3. Reverse Stress Testing

Instead of “What losses under X?”, asks: What shocks cause insolvency? Identifies vulnerabilities, guiding capital buffers and hedging overlays.

 

Summary

Quantitative risk management transforms survival into mathematical science.

  • Move beyond Gaussian assumptions with historical and Monte Carlo VaR.

  • Replace flawed VaR with Expected Shortfall and EVT.

  • Execute multi-factor stress tests to anticipate systemic shocks.

Together, these tools safeguard solvency across volatile global markets.

 
PreviousNext