Risk Management

Why Risk Management?

Risk management is the systematic identification, assessment, and mitigation of financial losses. It has evolved from basic position monitoring to sophisticated mathematical frameworks encompassing market, credit, operational, and liquidity risks. Modern risk management is essential for regulatory compliance, capital allocation, performance evaluation, and organizational survival.

The 2008 financial crisis demonstrated that inadequate risk management can lead to systemic failures, making it a cornerstone of modern finance for banks, hedge funds, asset managers, and corporations.

Types of Financial Risk

1. Market Risk

Definition: Risk of losses due to adverse price movements in financial markets.

Components:

Equity risk: Stock price volatility
Interest rate risk: Bond price sensitivity to rate changes
Currency risk: Foreign exchange rate fluctuations
Commodity risk: Price changes in raw materials
Volatility risk: Changes in implied volatility levels

2. Credit Risk

Definition: Risk of loss due to borrower default or credit deterioration.

Components:

Default risk: Probability of complete non-payment
Migration risk: Credit rating downgrades
Concentration risk: Large exposures to single borrowers
Counterparty risk: Derivatives and trading counterparts

3. Operational Risk

Definition: Risk of loss from inadequate systems, processes, people, or external events.

Categories:

Technology failures: System outages, cyber attacks
Human error: Trading mistakes, fraud
External events: Natural disasters, terrorism
Legal risk: Regulatory violations, litigation

4. Liquidity Risk

Definition: Risk of inability to meet funding obligations or liquidate positions.

Types:

Funding liquidity: Access to cash when needed
Market liquidity: Ability to trade without price impact
Asset-liability mismatch: Duration and currency mismatches

Value at Risk (VaR)

Definition

VaR: Maximum expected loss over a specific time horizon at a given confidence level.

\text{VaR}_\alpha = -F^{-1}(\alpha)

where $F$ is the cumulative distribution function of returns and $\alpha$ is the confidence level (e.g., 95%, 99%).

Parametric VaR

Assumptions: Portfolio returns are normally distributed.

Formula:

\text{VaR}_\alpha = \mu - \sigma \Phi^{-1}(\alpha)

where $\mu$ is expected return, $\sigma$ is volatility, and $\Phi^{-1}$ is the inverse normal CDF.

Portfolio VaR:

\text{VaR}_\alpha = \Phi^{-1}(\alpha) \sqrt{w^T \Sigma w}

for a portfolio with weights $w$ and covariance matrix $\Sigma$ .

Historical Simulation

Method:

Collect historical returns: Typically 250-500 days
Apply to current portfolio: Calculate P&L for each historical scenario
Sort outcomes: Rank from worst to best
Select percentile: VaR is the $\alpha$ -th percentile

Advantages:

Non-parametric: No distribution assumptions
Captures fat tails: Uses actual market data
Easy to understand: Intuitive methodology

Disadvantages:

Limited history: May miss rare events
Equal weighting: Recent events not emphasized
Computational cost: Requires full revaluation

Monte Carlo Simulation

Process:

Model specification: Choose stochastic process for risk factors
Parameter estimation: Calibrate model to market data
Simulation: Generate thousands of scenario paths
Portfolio valuation: Calculate P&L for each scenario
Statistical analysis: Determine VaR from simulated distribution

Advanced Techniques:

Importance sampling: Focus on tail scenarios
Variance reduction: Antithetic variables, control variates
Quasi-Monte Carlo: Low-discrepancy sequences

Expected Shortfall (CVaR)

Definition

Expected Shortfall: Average loss beyond VaR threshold.

\text{ES}_\alpha = \mathbb{E}[L | L \geq \text{VaR}_\alpha] = \frac{1}{1-\alpha} \int_\alpha^1 \text{VaR}_u du

Properties

Coherent Risk Measure: Satisfies all desirable properties:

Monotonicity: If $X \geq Y$ , then $\rho(X) \leq \rho(Y)$
Translation invariance: $\rho(X + c) = \rho(X) - c$
Homogeneity: $\rho(\lambda X) = \lambda \rho(X)$ for $\lambda \geq 0$
Subadditivity: $\rho(X + Y) \leq \rho(X) + \rho(Y)$

Computational Advantages:

Convex optimization: Easier to minimize than VaR
Differentiable: Enables gradient-based methods
Stable: Less sensitive to extreme outliers

Estimation

Historical Simulation:

\hat{\text{ES}}_\alpha = \frac{1}{n(1-\alpha)} \sum_{i: L_i \geq \hat{\text{VaR}}_\alpha} L_i

Parametric (Normal distribution):

\text{ES}_\alpha = \mu + \sigma \frac{\phi(\Phi^{-1}(\alpha))}{1-\alpha}

where $\phi$ is the standard normal PDF.

Stress Testing

Scenario Analysis

Historical Scenarios:

1987 Black Monday: 22% single-day equity decline
1998 LTCM Crisis: Credit spread widening, flight to quality
2008 Financial Crisis: Credit crunch, liquidity freeze
2020 COVID-19: Sudden volatility spike, correlation breakdown

Hypothetical Scenarios:

Interest rate shocks: Parallel and non-parallel shifts
Credit spread widening: Uniform or sector-specific
Equity market crashes: Regional or global selloffs
Currency crises: Major devaluations

Sensitivity Analysis

Greeks for Options:

Delta: $\frac{\partial V}{\partial S}$ (price sensitivity)
Gamma: $\frac{\partial^2 V}{\partial S^2}$ (convexity)
Theta: $\frac{\partial V}{\partial t}$ (time decay)
Vega: $\frac{\partial V}{\partial \sigma}$ (volatility sensitivity)
Rho: $\frac{\partial V}{\partial r}$ (interest rate sensitivity)

Factor Sensitivities:

\Delta P \approx \sum_{i=1}^n \frac{\partial P}{\partial F_i} \Delta F_i + \frac{1}{2} \sum_{i=1}^n \sum_{j=1}^n \frac{\partial^2 P}{\partial F_i \partial F_j} \Delta F_i \Delta F_j

Reverse Stress Testing

Approach: Determine scenarios that would cause unacceptable losses.

Method:

Define failure threshold: Maximum tolerable loss
Optimize scenarios: Find scenarios achieving this loss
Assess plausibility: Evaluate scenario likelihood
Mitigation strategies: Develop contingency plans

Model Risk

Definition

Model Risk: Risk of losses due to incorrect model specification, calibration, or application.

Sources

Specification Risk:

Wrong functional form: Linear vs. non-linear relationships
Missing variables: Omitted risk factors
Structural breaks: Model relationships change over time

Calibration Risk:

Parameter uncertainty: Estimation error in model inputs
Data quality: Incomplete, biased, or stale data
Sample selection: Non-representative historical periods

Implementation Risk:

Programming errors: Bugs in model code
Numerical issues: Convergence, stability problems
Data feed errors: Incorrect market data inputs

Model Validation

Quantitative Tests:

Backtesting: Compare predictions to realized outcomes
Cross-validation: Out-of-sample performance testing
Sensitivity analysis: Parameter stability assessment

Qualitative Review:

Model documentation: Theory, assumptions, limitations
Code review: Programming standards, testing procedures
Usage monitoring: Appropriate application scope

Model Risk Mitigation

Model diversification: Use multiple models for critical decisions Champion-challenger: Compare new models to existing ones Model limits: Restrict model usage to validated domains Override procedures: Human judgment in extreme conditions

Liquidity Risk Management

Funding Liquidity

Liquidity Coverage Ratio (LCR):

\text{LCR} = \frac{\text{High Quality Liquid Assets}}{\text{Net Cash Outflows over 30 days}} \geq 100\%

Net Stable Funding Ratio (NSFR):

\text{NSFR} = \frac{\text{Available Stable Funding}}{\text{Required Stable Funding}} \geq 100\%

Market Liquidity

Bid-Ask Spread Analysis:

\text{Spread} = \frac{\text{Ask} - \text{Bid}}{\text{Mid-point}}

Market Impact Models:

\text{Impact} = \alpha \times \left(\frac{\text{Trade Size}}{\text{Daily Volume}}\right)^\beta

Liquidation Time:

T = \frac{\text{Position Size}}{\text{Participation Rate} \times \text{Average Daily Volume}}

Contingent Liquidity

Stressed scenarios: Model liquidity during crisis periods Contingency funding: Emergency funding sources Asset fire sale: Impact of forced liquidations

Regulatory Capital

Basel III Framework

Common Equity Tier 1 (CET1):

\text{CET1 Ratio} = \frac{\text{CET1 Capital}}{\text{Risk-Weighted Assets}} \geq 4.5\%

Total Capital Ratio:

\text{Total Capital Ratio} = \frac{\text{Total Capital}}{\text{Risk-Weighted Assets}} \geq 8\%

Leverage Ratio:

\text{Leverage Ratio} = \frac{\text{Tier 1 Capital}}{\text{Total Exposure}} \geq 3\%

Risk-Weighted Assets

Standardized Approach: Regulatory risk weights by asset class Internal Ratings-Based: Bank's own risk models for:

PD: Probability of default
LGD: Loss given default
EAD: Exposure at default

Capital Formula:

K = (LGD \times N(\sqrt{\frac{1}{1-R}} N^{-1}(PD) + \sqrt{\frac{R}{1-R}} N^{-1}(0.999)) - PD \times LGD) \times M

where $R$ is correlation and $M$ is maturity adjustment.

Operational Risk

Measurement Approaches

Basic Indicator Approach: 15% of gross income Standardized Approach: Different percentages by business line Advanced Measurement Approach: Internal models using:

Loss Distribution Approach: Frequency × severity modeling
Scenario-based models: Expert judgment scenarios
Scorecard approaches: Risk indicators and controls

Key Risk Indicators (KRIs)

Technology:

System downtime: Hours of outages per month
Failed trades: Percentage of failed settlements
Cyber incidents: Number of security breaches

Human:

Employee turnover: Staff retention rates
Training completeness: Percentage trained on procedures
Error rates: Operational mistakes per transaction

Business Continuity

Disaster recovery: Backup systems and data centers Crisis management: Communication and decision procedures Stress scenarios: Operations during extreme events

Risk Aggregation

Correlation Modeling

Linear correlation: Pearson correlation coefficient Rank correlation: Spearman's rho for non-linear relationships Tail dependence: Copula models for extreme events

Diversification Benefits

Portfolio variance:

\sigma_p^2 = \sum_{i=1}^n w_i^2 \sigma_i^2 + 2\sum_{i<j} w_i w_j \sigma_i \sigma_j \rho_{ij}

Diversification ratio:

DR = \frac{\sum_{i=1}^n w_i \sigma_i}{\sigma_p}

Risk Budgeting

Risk contribution:

RC_i = w_i \frac{\partial \sigma_p}{\partial w_i} = w_i \frac{(\Sigma w)_i}{\sigma_p}

Component VaR:

\text{CVaR}_i = \text{VaR}_p \times \frac{RC_i}{\sigma_p^2}

Technology and Infrastructure

Risk Management Systems

Data management: Market data, positions, transactions Calculation engines: VaR, stress testing, sensitivity analysis Reporting platforms: Dashboards, alerts, regulatory reports Integration: Trading systems, accounting, regulatory reporting

Real-Time Monitoring

Position monitoring: Live P&L and risk metrics Limit management: Automated limit checking and alerts Exception handling: Workflow for limit breaches Kill switches: Emergency stop mechanisms

Data Quality

Validation rules: Automated checks for data integrity Reconciliation: Comparison across multiple sources Error handling: Exception workflows and corrections Audit trails: Complete history of data changes

Connection to Other Topics

Risk management integrates many quantitative concepts:

Built on probability distributions and statistical methods
Uses stochastic processes for scenario modeling
Applies portfolio optimization for risk allocation
Connects to credit risk models and interest rate models
Foundation for algorithmic trading risk controls
Links to option pricing for derivatives risk
Enables regulatory compliance and capital management