Unit
- Definition: Value at Risk (VaR) is a statistical measure that estimates the maximum potential loss an investment portfolio could suffer over a specified time horizon at a given confidence level. It provides a single number summarizing downside risk, enabling comparison across different assets and portfolios.
- Interpretation: A daily VaR of $1 million at 95% confidence means there is a 5% chance that the portfolio will lose more than $1 million in one day. The metric does not indicate how large the loss might be beyond the VaR threshold, which is a limitation.
- Purpose: VaR is used by financial institutions to set risk limits, allocate capital, and report risk exposure to regulators and stakeholders. It helps in assessing whether risk levels align with the firm's risk appetite.
- Historical Context: VaR gained prominence after the 1987 stock market crash and the 1990s derivatives scandals. The 1994 JPMorgan RiskMetrics publication popularized the concept, leading to its adoption as a standard risk tool.
- Key Components: VaR depends on three parameters: the portfolio composition, the time horizon (e.g., 1 day, 10 days), and the confidence level (e.g., 95%, 99%). These parameters are chosen based on the use case and regulatory guidelines.
- Common Confidence Levels: 95% and 99% are widely used. A 99% confidence level yields a higher VaR than 95% because it captures a larger tail of the distribution. Regulators often require 99% VaR for capital adequacy.
- Time Horizon Considerations: Shorter horizons (e.g., 1 day) are used for trading portfolios, while longer horizons (e.g., 10 days) are typical for less liquid assets. The horizon affects the magnitude of potential losses due to compounding.
- Limitations: VaR assumes normal market conditions and ignores extreme tail events. It is not subadditive, meaning the VaR of a combined portfolio can be greater than the sum of individual VaRs, contradicting diversification benefits.
- Calculation Methods: VaR can be calculated using three primary approaches. Historical simulation uses past returns, parametric (variance-covariance) assumes normal distribution, and Monte Carlo simulates random price paths. Each method has trade offs between speed, accuracy, and assumptions about return distributions.
- Backtesting: Firms regularly backtest VaR models by comparing predicted losses with actual outcomes. If losses exceed VaR more frequently than the confidence level implies, the model is considered inaccurate. Backtesting helps validate model assumptions and prompts recalibration when violations are too numerous.
- Conditional VaR (CVaR): Also known as Expected Shortfall, CVaR measures the average loss beyond the VaR threshold. It addresses VaR’s limitation of ignoring tail severity. For a 95% confidence level, CVaR tells you the expected loss in the worst 5% of cases, providing a more complete risk picture.
- Regulatory Framework: Under Basel II and III, banks must calculate VaR for market risk capital requirements. The standard is a 99% confidence level over a 10 day holding period, scaled from daily VaR. Regulators also require stress testing and backtesting to supplement VaR based capital charges.
Risk Assessment
MEDIUM RISK
VaR is a foundational risk measure but must be used alongside other tools. Misinterpretation can lead to underestimation of tail risk. Practitioners should recognize that VaR captures only the threshold loss, not the severity of losses beyond it. The medium risk level reflects that while VaR is effective for everyday monitoring, heavy reliance on it without stress testing can expose the portfolio to black swan events.
KPI Card
Typical VaR (Illustrative)
$1.2M
+15% YoY (due to volatility increase)
This example shows how VaR can change over time. For a $100M portfolio at 95% confidence over 1 day, a VaR of $1.2M implies that losses exceed $1.2M only 5% of days. Monitoring VaR trends helps identify increasing risk.
Quick Revision
- VaR estimates maximum loss over a period at a given confidence level.
- It is a single number summarizing downside risk, widely used in finance.
- Key inputs: portfolio composition, time horizon, confidence level.
- VaR has limitations, including ignoring tail risk and lack of subadditivity.