While Value-at-Risk (VaR) is widely used in risk management, it has several limitations that make it insufficient as a standalone risk measure. Below are the key drawbacks, categorized for a structured response in an interview.
Does Not Capture Tail Risk (Extreme Losses Beyond VaR)
VaR only tells us the threshold loss at a given confidence level but does not indicate how bad losses can be beyond that level. for example: A 95% one-day VaR of $1M means there’s a 5% chance the loss exceeds $1M, but it does not quantify how much the loss could actually be—whether it's $1.1M, $5M, or $50M.
Alternative: this is why Expected Shortfall (Conditional VaR or CVaR) is preferred, as it estimates the average loss beyond VaR.
Assumes Normal Distribution (In Parametric Approach): the variance-covariance method (parametric VaR) assumes asset returns follow a normal distribution. However, financial markets often exhibit fat tails and skewness, meaning large losses occur more frequently than a normal distribution would predict. for example: the 2008 Financial Crisis involved extreme losses that normal VaR models significantly underestimated.
Ignores Liquidity Risk: VaR assumes that positions can be liquidated instantly at current market prices. In reality:
During market stress, liquidity dries up, making it impossible to exit positions at quoted prices.
Bid-ask spreads widen, and large trades can move the market against the seller.
for example, Long-Term Capital Management (LTCM) Collapse (1998)
the hedge fund's highly leveraged positions became illiquid, leading to losses far exceeding VaR estimates.
Does Not Capture Correlation Breakdown During Crises: VaR relies on historical data to estimate correlations between assets, assuming they remain stable. However, in times of financial stress, asset correlations tend to increase (everything sells off together), leading to losses greater than predicted.
for example,
In a normal market, equity and bonds might have negative correlation (stocks fall, bonds rise).
During crises, correlations converge to 1 as investors liquidate all assets for cash, leading to simultaneous drawdowns.
Highly Sensitive to the Lookback Period: VaR is calculated based on historical data, but the chosen time period significantly affects results. for example:
If you use data from a calm market period, VaR underestimates risk.
If you use data from a volatile period, VaR overestimates risk.
this makes VaR highly sensitive to recent market conditions and can give misleading results.
Can Be Manipulated (Subject to Model Risk):
Different choices of time horizon, confidence level, or calculation method can lead to vastly different VaR values. Financial institutions may tweak assumptions to show a lower VaR to meet regulatory requirements, leading to regulatory arbitrage.
for example, a bank calculating 1-day 99% VaR might report a low number, but switching to 10-day 99% VaR could reveal much higher risk.
Alternative: Regulators prefer Expected Shortfall (ES) under Basel III/FRTB, as it is harder to manipulate than VaR.
Does Not Measure Frequency of Losses
VaR tells us the maximum expected loss at a confidence level, but it does not indicate how often smaller losses occur. A portfolio with a low VaR may still have frequent smaller losses, which can erode capital over time.
Alternative: Expected Shortfall and drawdown analysis provide better insights into risk exposure.
Does Not Differentiate Between Small and Large Portfolios
VaR is often expressed as an absolute dollar amount ($X million loss), which does not account for portfolio size. for example,
A $10 million VaR is a huge risk for a $100 million portfolio (10% exposure).
The same $10M VaR is minor for a $10 billion portfolio (0.1% exposure).
Alternative: Using VaR as a percentage of total portfolio value improves comparability.
Not Additive Across Portfolios
VaR is not sub-additive, meaning the total VaR of a combined portfolio is not necessarily the sum of individual VaRs. This makes it difficult to aggregate risk across multiple desks, assets, or subsidiaries.
Alternative: Expected Shortfall is sub-additive, making it a more coherent risk measure.
How to Answer in an Interview
Start with a Strong Statement:
"While VaR is a widely used risk measure, it has several limitations that make it insufficient on its own."
Explain the Key Drawbacks Clearly:
First, VaR does not capture tail risk, meaning it does not measure the size of losses beyond the threshold. This is why Expected Shortfall is preferred.
Second, it assumes normal distribution in the parametric method, but market returns often have fat tails, leading to underestimated risk.
Third, it does not account for liquidity risk—VaR assumes we can liquidate positions at market prices, which is not always possible in stressed conditions.
Mention Alternative Measures:
Because of these limitations, risk managers often use Expected Shortfall, stress testing, and liquidity risk analysis alongside VaR to get a more complete picture of risk.
Keep It Concise but Technical:
If the interviewer asks for an in-depth explanation, discuss correlation breakdown, model risk, and regulatory implications (Basel III/FRTB).
Your Interview Answer
While Value-at-Risk is a useful measure for quantifying potential losses, it has several limitations. The most critical drawback is that VaR does not capture extreme tail losses—meaning it only tells us the threshold loss at a confidence level but not how bad losses can be beyond that.
Additionally, parametric VaR assumes a normal distribution of returns, which often underestimates risk because financial markets exhibit fat tails.
VaR also does not account for liquidity risk, correlation breakdowns during crises, or model risk, making it susceptible to misinterpretation.
Furthermore, VaR is sensitive to the chosen historical lookback period and is not additive across portfolios, making it difficult to aggregate risk.
Due to these limitations, Expected Shortfall (Conditional VaR) is increasingly used in regulatory frameworks such as Basel III and FRTB.
Comments