Value at Risk: Measuring and Managing Portfolio Risk

Value at Risk (VaR) is a risk measure that estimates a portfolio loss threshold over a specified time horizon and at a chosen confidence level. It is commonly used by financial institutions to summarise and compare market-risk exposure across portfolios and positions.

VaR does not represent the maximum possible loss. Its calculation and interpretation depend on factors including the methodology, underlying data, assumptions, time horizon, and confidence level.

What Is Value at Risk (VaR)?

Value at Risk expresses a loss threshold for a portfolio over a defined time horizon and at a specified confidence level.

Three components are important when interpreting a VaR figure:

• Loss threshold: The monetary amount or percentage represented by the VaR calculation.

• Time horizon: The period covered by the calculation, such as one trading day or 10 trading days.

• Confidence level: The probability level associated with the calculation, commonly expressed as 95% or 99%, depending on its application.

For example, consider a portfolio with a one day 95% VaR of $1 million. Under the model's assumptions, this indicates a 95% probability that the portfolio's loss over one day does not exceed $1 million. Losses exceeding $1 million correspond to the remaining 5% of the modeled distribution.

The VaR figure does not describe the size of losses beyond that threshold.

Read also: The Importance of Risk Management in Volatile Market

How Is Value at Risk Calculated?

VaR calculations use different methodologies. Common approaches include:

• Historical simulation: Applies historical market movements to a portfolio to estimate the distribution of potential outcomes.

• Parametric VaR: Uses statistical assumptions about market returns and portfolio exposures to estimate the loss threshold.

• Monte Carlo simulation: Uses simulated market scenarios to estimate a distribution of portfolio outcomes.

Different methodologies, datasets and assumptions can produce different VaR estimates for the same portfolio. The methodology and assumptions underlying a VaR figure are therefore relevant to its interpretation.

Limitations of Value at Risk

One limitation of VaR is that it identifies a loss threshold at a specified confidence level without measuring the severity of losses beyond that threshold.

For example, two portfolios with similar VaR figures can have different loss distributions beyond their respective VaR thresholds.

Expected shortfall (ES) provides a different measure of tail risk by estimating the average loss in the tail of the loss distribution beyond a specified confidence threshold. Under the Basel market-risk framework, expected shortfall is incorporated into the internal models approach for market-risk measurement.

Other limitations associated with VaR include:

• Model dependence: Methodologies, statistical assumptions, lookback periods, and datasets affect the resulting VaR calculation.

• Historical-data limitations: Models based on historical observations reflect the market conditions contained in the selected dataset.

• Changes in correlations and volatility: Correlations and volatility observed during stressed market conditions can differ from observations during less volatile periods.

• Tail risk: A VaR figure identifies a threshold but does not describe the magnitude of losses beyond that threshold.

• Model risk: Differences in model design, assumptions, and inputs can result in different risk estimates for the same or similar exposures.

  
  

These characteristics are among the reasons VaR appears alongside other measures within institutional risk-management frameworks.

How Institutions Use Value at Risk in Portfolio Risk Management

VaR is used within institutional risk-management frameworks as a standardized measure of market-risk exposure.

Risk functions calculate VaR at different levels, including individual portfolios, trading desks, strategies, asset classes and aggregated portfolios. This provides a common measure for comparing market risk exposures across different parts of an institution.

Depending on an institution's risk-management framework, VaR is also incorporated into areas such as:

• Risk monitoring

• Internal risk limits.

• Risk reporting.

• Exposure monitoring.

• Model backtesting

Value at Risk and Backtesting

Backtesting compares model-generated VaR estimates with observed portfolio outcomes.

The comparison provides information about how frequently actual losses exceed the VaR threshold relative to the frequency implied by the selected confidence level. Backtesting therefore forms part of the processes used to assess the performance of VaR models.

Value at Risk Alongside Other Risk Measures

Institutional risk-management frameworks incorporate a range of measures and controls rather than representing portfolio risk through a single metric.

Alongside VaR, these frameworks include measures and processes such as:

• Stress testing.

• Scenario analysis.

• Sensitivity measures.

• Expected shortfall.

• Exposure and risk limits.

• Backtesting.

• Model validation.

Each addresses different aspects of market risk

The Basel market-risk framework, for example, incorporates expected shortfall within its internal models approach. Expected shortfall provides information about losses in the tail of the distribution beyond the relevant threshold.

Read also: Institutional Investor Risk Management: Modern Strategies

Conclusion

Value at Risk is a widely used risk measure that estimates a portfolio loss threshold over a specified time horizon and at a stated confidence level.

The figure does not represent a portfolio's maximum possible loss. Its interpretation depends on the methodology, data, and assumptions underlying the calculation.

Within institutional risk-management frameworks, VaR is used alongside measures and processes including stress testing, scenario analysis, expected shortfall, risk limits, backtesting and model validation. Together, these tools provide different perspectives on portfolio and market risk exposures.