详细指南即将推出
我们正在为Conditional VaR (CVaR/ES)编写全面的教育指南。请尽快回来查看逐步解释、公式、真实案例和专家提示。
Conditional Value at Risk (CVaR), also known as Expected Shortfall (ES) or Expected Tail Loss (ETL), is a risk measure that quantifies the average loss in the worst-case scenarios that exceed the Value at Risk (VaR) threshold. While VaR answers 'what is the maximum loss I expect not to exceed at a given confidence level,' CVaR answers 'given that I have exceeded the VaR threshold, what is the average loss I should expect?' This makes CVaR a superior risk measure for capturing tail risk — the catastrophic losses that happen infrequently but with devastating impact. For example, if the 95% 1-day VaR of a portfolio is $1 million, it means there is a 5% chance of losing more than $1 million in a day. The 95% CVaR would then be the average of all losses in that worst 5% of scenarios — perhaps $1.8 million. The CVaR is always greater than or equal to VaR at the same confidence level, and provides critical information about the severity (not just the probability) of extreme losses. CVaR has become the preferred risk measure in modern risk management and regulatory frameworks for several reasons. First, it is coherent: unlike VaR, CVaR satisfies the mathematical property of subadditivity (the CVaR of a combined portfolio is no greater than the sum of individual CVaRs), meaning diversification is always rewarded. VaR lacks this property. Second, CVaR gives more information about the tail of the loss distribution. Third, CVaR is convex and can be optimized using linear programming, making it practical for portfolio optimization. The Basel Committee on Banking Supervision (BCBS) formally shifted the market risk framework from VaR to Expected Shortfall in the Fundamental Review of the Trading Book (FRTB), which took full effect in 2023. Under FRTB, banks must calculate ES at a 97.5% confidence level (equivalent in tail coverage to 99% VaR under normal distributions), capturing more tail risk. This regulatory shift acknowledges that VaR systematically underestimates risk during market crises, as demonstrated by the 2008 Global Financial Crisis and subsequent stress periods. CVaR can be calculated analytically (for normally distributed returns), from historical simulation (averaging observed losses in the tail), or using Monte Carlo simulation (averaging simulated tail losses). Each method has strengths and weaknesses, and sophisticated risk systems typically use all three as cross-checks.
Parametric CVaR (normal): CVaR_α = −μ + σ × φ(Φ⁻¹(α)) / (1−α) Historical CVaR: Average of all observed losses exceeding VaR_α Parallel: CVaR = E[Loss | Loss > VaR_α]
- 1Specify confidence level (α): typically 95% or 99% for internal risk management; 97.5% for Basel III FRTB regulatory capital.
- 2Collect portfolio return data: historical returns (historical simulation) or model parameters μ and σ (parametric).
- 3Calculate VaR at level α: the loss exceeded with probability (1−α). For normal: VaR = −μ + σ × Φ⁻¹(α).
- 4For historical CVaR: sort all historical losses from worst to best; average the losses in the worst (1−α)% of observations.
- 5For parametric CVaR (normal): CVaR = −μ + σ × [φ(Φ⁻¹(α)) / (1−α)], where φ is the standard normal PDF and Φ⁻¹ is the inverse CDF.
- 6For Monte Carlo CVaR: simulate thousands of portfolio return scenarios; sort by loss; average the worst (1−α)% of simulated outcomes.
- 7Report CVaR alongside VaR and the ES/VaR ratio. Compare to capital reserves and risk limits. Under Basel FRTB, ES replaces VaR as the primary regulatory capital measure.
CVaR is 15% larger than VaR — captures severity of extreme losses
For a normal distribution: VaR₉₉% = σ × Φ⁻¹(0.99) = 1.5% × 2.326 = 3.489% → $348,900. CVaR₉₉% = σ × φ(2.326) / 0.01 = 1.5% × (0.02665 / 0.01) = 1.5% × 2.665 = 3.998% → $399,800. CVaR of $400K represents the average daily loss in the worst 1% of days. This means on the very worst days, the portfolio expects to lose $400,000 on average — not just the $348,900 VaR threshold. Risk managers use CVaR to size capital buffers for tail events.
Historical tail is fatter than normal — CVaR/VaR ratio=1.47x
With 2,520 observations at 95% confidence: worst 5% = 126 observations. Sort all returns from worst to best; the 126th worst is the VaR cutoff (loss = 1.5% × $50M = $750K). Average all 126 worst losses: if the average worst-5% loss is 2.2%, CVaR = 2.2% × $50M = $1,100,000. The CVaR/VaR ratio of 1.47 reflects fat tails in actual equity returns — real distributions have more extreme events than normal distribution predicts. This is why historical simulation often produces higher CVaR than parametric methods.
FRTB uses ES at 97.5% over stressed historical scenarios — replacing old 99% VaR
Under Basel III FRTB (effective 2023), market risk capital for trading books uses Expected Shortfall at 97.5% confidence over a 10-day liquidity horizon, calibrated to a stressed historical period. A 97.5% ES under normality is approximately equal to a 99% VaR in tail area, but ES captures the magnitude of losses beyond the threshold. Banks must identify the worst one-year stress period in their history and calculate ES using that data. Capital = ES_stressed × liquidity adjustment factors for different asset classes.
CVaR is subadditive — VaR is not (VaR fails this property)
The combined portfolio CVaR of $650K is less than the sum of individual CVaRs ($500K + $400K = $900K), demonstrating CVaR's subadditivity property. The $250K difference is the diversification benefit — assets A and B don't perfectly co-crash, so tail losses are partially offset. VaR does not always exhibit this property (VaR of combined can exceed the sum in some cases), which is why VaR is not a coherent risk measure while CVaR is. This mathematical coherence makes CVaR preferable for portfolio optimization and regulatory capital.
Bank trading book regulatory capital (Basel III FRTB), enabling practitioners to make well-informed quantitative decisions based on validated computational methods and industry-standard approaches, which requires precise quantitative analysis to support evidence-based decisions, strategic resource allocation, and performance optimization across diverse organizational contexts and professional disciplines
Hedge fund risk monitoring and investor reporting, helping analysts produce accurate results that support strategic planning, resource allocation, and performance benchmarking across organizations, where accurate numerical computation is essential for producing reliable outputs that inform planning, evaluation, and continuous improvement processes in both corporate and individual settings
Insurance company reserve adequacy under Solvency II, allowing professionals to quantify outcomes systematically and compare scenarios using reliable mathematical frameworks and established formulas, demanding systematic calculation approaches that translate raw input data into actionable insights for stakeholders who depend on quantitative rigor in their daily professional activities
Portfolio optimization with tail risk constraints, supporting data-driven evaluation processes where numerical precision is essential for compliance, reporting, and optimization objectives, necessitating robust computational methods that deliver consistent and verifiable results suitable for reporting, auditing, and long-term trend analysis in professional environments
Pension fund asset-liability risk management, which requires precise quantitative analysis to support evidence-based decisions, strategic resource allocation, and performance optimization across diverse organizational contexts and professional disciplines
Under a Student's t-distribution with 4 df, ES/VaR ratios are 50–100% higher than normal, meaning parametric normal CVaR significantly underestimates actual tail risk. Always supplement parametric CVaR with historical simulation."}. Professionals working with conditional var should be especially attentive to this scenario because it can lead to misleading results if not handled properly. Always verify boundary conditions and cross-check with independent methods when this case arises in practice.
Professionals working with conditional var should be especially attentive to this scenario because it can lead to misleading results if not handled properly. Always verify boundary conditions and cross-check with independent methods when this case arises in practice.
Professionals working with conditional var should be especially attentive to this scenario because it can lead to misleading results if not handled properly. Always verify boundary conditions and cross-check with independent methods when this case arises in practice.
| Confidence Level | VaR Multiplier (z) | ES Multiplier | ES/VaR Ratio | Basel Application |
|---|---|---|---|---|
| 90% | 1.282 | 1.755 | 1.370 | Internal risk limit (common) |
| 95% | 1.645 | 2.063 | 1.254 | Standard risk reporting |
| 97.5% | 1.960 | 2.338 | 1.193 | Basel III FRTB primary metric |
| 99% | 2.326 | 2.665 | 1.145 | Legacy Basel II/2.5 standard |
| 99.5% | 2.576 | 2.892 | 1.123 | Insurance / Solvency II |
| 99.9% | 3.090 | 3.368 | 1.090 | Economic capital / extreme risk |
Why is CVaR considered superior to VaR?
CVaR is superior to VaR for several interconnected reasons. First, CVaR is coherent: it satisfies subadditivity (diversification always helps), while VaR does not. Second, CVaR provides information about loss severity beyond the threshold, not just the probability of exceeding it. VaR says 'you'll lose no more than X with probability α' — but gives no information about how bad losses are when they do exceed X. Third, CVaR is convex and can be incorporated into portfolio optimization frameworks using linear programming. Fourth, CVaR is more sensitive to the shape of the tail — fat-tailed distributions produce materially higher CVaR, giving an honest picture of tail risk that VaR can obscure.
What is the Basel III FRTB requirement for Expected Shortfall?
The Fundamental Review of the Trading Book (FRTB), Basel III's market risk framework, replaced 99% VaR with 97.5% Expected Shortfall as the primary capital metric. Key requirements include: ES must be calculated using a 10-business-day liquidity horizon (not the old 10-day scaled from 1-day); different asset classes have different liquidity horizons (from 10 days for large-cap equities to 120 days for some credit products); ES must be calibrated to a stressed historical period; and banks must also calculate a non-stressed ES for comparison. The framework took effect for large internationally active banks in 2023, with Basel III full implementation completed through 2025–2028 in various jurisdictions.
What is the relationship between VaR and CVaR at different confidence levels?
For normally distributed returns, the ES/VaR ratio is a fixed function of confidence level. At 95%, ES is approximately 25% higher than VaR. At 99%, ES is approximately 14% higher than VaR. At 99.9%, ES is about 8% higher. However, for fat-tailed (leptokurtic) distributions — which better describe actual financial returns — the ratio is substantially higher. A Student's t-distribution with 4 degrees of freedom produces ES/VaR ratios 50–100% higher than the normal case. This is why historical or Monte Carlo simulation methods typically produce higher CVaR than parametric normal models.
How does CVaR change during market stress periods?
CVaR increases dramatically during stress periods because: (1) portfolio volatility rises, shifting the entire loss distribution rightward; (2) correlations between assets increase toward 1.0 during crises, reducing diversification benefits; (3) return distributions become more negatively skewed and fat-tailed, amplifying tail losses beyond what volatility alone would predict. During the 2008 crisis, typical daily VaR estimates were breached 10–20 times more frequently than the implied 1% probability — demonstrating that historical CVaR estimates calibrated to tranquil periods severely understate tail risk in crises. This is why FRTB requires stressed-period calibration.
Can CVaR be used for portfolio optimization?
Yes, and this is one of CVaR's major practical advantages over VaR. Rockafellar and Uryasev (2000) showed that CVaR minimization can be formulated as a linear program, making it computationally tractable even for large portfolios. The mean-CVaR portfolio optimization framework minimizes CVaR for a given expected return (analogous to mean-variance Markowitz optimization), but incorporates tail risk into the objective function. This produces portfolios that are more robust to extreme events compared to mean-variance optimal portfolios, at a modest cost in expected return. Many institutional risk managers now use mean-CVaR optimization as their primary portfolio construction framework.
What is backtesting for CVaR and why is it difficult?
Backtesting VaR is straightforward: count how many days actual losses exceeded the VaR estimate and compare to the expected frequency (e.g., 1% for 99% VaR). CVaR backtesting is more challenging because CVaR is an average conditional on tail events — and tail events are rare, making statistical inference difficult with limited data. Tests include: (1) conditional coverage tests — do losses beyond VaR average to the predicted CVaR? (2) Acerbi-Szekely spectral tests — comparing the entire tail distribution. Because tail events are rare, CVaR backtests have low statistical power and require many years of data to detect even moderate underestimation. This is an active area of research in quantitative risk management.
How is CVaR different from stress testing?
CVaR is a statistical estimate of the average tail loss given the current distributional model. Stress testing examines specific worst-case scenarios — often historical crises (2008 GFC, COVID-19 crash, 1987 Black Monday) or hypothetical events (geopolitical shock, central bank policy reversal, sector collapse). CVaR is model-dependent and captures average tail behavior under the modeled distribution; stress tests capture specific scenario losses regardless of their statistical probability. Best practice combines both: CVaR for ongoing daily risk monitoring and capital allocation, stress testing for scenario-specific preparedness and board-level risk communication.
专业提示
Always report VaR and CVaR together, along with the CVaR/VaR ratio. A ratio much above 1.3–1.5 at 99% confidence signals fat tails in the return distribution and warns that the VaR underestimates the severity of tail events significantly.
你知道吗？
CVaR as a formal risk measure was introduced by Rockafellar and Uryasev in their 2000 paper 'Optimization of Conditional Value-at-Risk' in the Journal of Risk. The same paper showed how CVaR could be computed via a simple linear programming problem, making portfolio CVaR minimization practically feasible for the first time — a major advance in quantitative risk management.
参考资料
- ›Rockafellar & Uryasev (2000): Optimization of Conditional Value-at-Risk, Journal of Risk
- ›Basel Committee: Minimum Capital Requirements for Market Risk (FRTB, 2019)
- ›McNeil, Frey & Embrechts: Quantitative Risk Management (2nd ed.), Princeton University Press
- ›Investopedia: Conditional Value at Risk (CVaR)