Value at Risk (VaR): Definition, Methods, Calculation and CFP Exam Notes

Every investor knows that portfolios can lose money. But knowing that is different from knowing how much. An investor or fund manager who says “this portfolio could lose badly in a downturn” is stating the obvious. One who says “there is a 5% probability this portfolio will lose more than Rs 15 lakh over the next month” is saying something precise and actionable.

That precision is what Value at Risk (VaR) provides. It converts vague notions of risk into a specific number with a specific probability attached to it. For institutional investors, banks, regulators, and increasingly for retail financial planners, VaR is the standard tool for communicating and managing portfolio risk.

For CFP exam candidates in India, VaR is tested under Investment Planning (Module 4) as part of the quantitative risk measurement toolkit. It is also directly relevant to SEBI’s regulatory framework: SEBI mandates VaR-based margins on NSE and BSE for all equity and derivative positions. Understanding how VaR works, how it is calculated using three distinct methods, and where it fits alongside other risk measures is essential for the exam and for professional practice.

1. What Is Value at Risk?

Value at Risk (VaR) is a statistical measure that estimates the maximum potential loss of a portfolio or investment over a specified time period at a given level of confidence under normal market conditions.

Value at Risk is a widely used risk measurement tool that estimates the maximum potential loss an investment or portfolio could face over a specific time period under normal market conditions. It is usually expressed as a monetary value or percentage and is calculated for a given confidence level, such as 95% or 99%.

In simple terms, VaR answers one question: “What is the worst I can expect to lose, with a given degree of certainty, over a defined period?”

The answer to that question has three components: a loss amount, a time period, and a probability. All three must be specified for a VaR statement to be meaningful.

VaR represents the maximum expected loss over a specified period at a certain confidence interval such as 95% or 99%. For instance, a one-day 95% VaR of Rs 1 lakh indicates there is only a 5% chance of losing more than Rs 1 lakh in a day under normal market conditions. VaR informs financial firms, regulators, and investors about capital reserves required to cover potential losses.

2. The Three Key Components of VaR

Every VaR calculation has three inputs that must be explicitly defined:

1. Time Horizon

The time period over which potential losses are evaluated. Common choices are one day (used by traders and for margining), one week, one month, or one year. The time horizon should match the investment’s liquidity and the purpose of the analysis.

2. Confidence Level

The probability that the actual loss will not exceed the VaR estimate. The most common confidence levels are 95% and 99%.

A 95% confidence level means there is a 5% probability (or 1 in 20 chance) that the actual loss will exceed the VaR figure. A 99% confidence level means there is only a 1% probability (or 1 in 100 chance) that the loss will exceed the VaR.

Higher confidence levels produce higher VaR figures because they attempt to capture a more extreme tail of the loss distribution.

3. Loss Amount

The maximum expected loss in absolute rupee terms or as a percentage of portfolio value at the chosen confidence level and time horizon.

3. How to Read a VaR Statement

“The 10-day 99% VaR of this portfolio is Rs 25 lakh.”

This statement means:

• Over any 10-day period
• There is a 99% probability that the portfolio will not lose more than Rs 25 lakh
• Equivalently, there is a 1% probability that the loss will exceed Rs 25 lakh

It does not say what the loss will be in that 1% scenario. That is the limitation that Conditional VaR (CVaR) addresses.

“The one-day 95% VaR of this fund is 2.068%.”

This means there is a 5% chance that on any given day, the fund could lose more than 2.068% of its value.

4. Method 1: Parametric VaR (Variance-Covariance Method)

The parametric method, also called the variance-covariance method, assumes that investment returns follow a normal distribution. It uses the mean return, standard deviation, and a Z-score corresponding to the chosen confidence level.

This analytical method assumes normally distributed returns. It uses mean, standard deviation, and confidence level using the Z-score to compute risk. For diversified portfolios, the covariance matrix between assets is employed to improve accuracy.

Parametric VaR Formula:

VaR = Portfolio Value x Standard Deviation x Z-score x √Time

Z-scores for common confidence levels:

• 90% confidence: Z = 1.28
• 95% confidence: Z = 1.65
• 99% confidence: Z = 2.33

Key characteristics:

• Fast and computationally simple
• Requires only mean and standard deviation as inputs
• Assumes symmetric, normally distributed returns
• Less accurate for portfolios with options or non-linear instruments
• Works well for large, diversified equity portfolios

This approach assumes that a portfolio follows a normal distribution. The fund manager first estimates a mean return using models such as the capital asset pricing model, historical averages, or multifactor models. The manager also estimates the asset’s standard deviation. Once the confidence level is decided, either 95% or 99%, the calculation produces the VaR figure.

5. Method 2: Historical Simulation VaR

The historical simulation method uses actual past return data to estimate future loss potential. It does not assume any particular distribution of returns. Instead, it ranks historical returns from worst to best and identifies the loss at the chosen percentile.

This model uses actual historical return data, reordering past losses to determine the loss at a given percentile. It avoids assumptions about return distributions and offers an empirical approach to estimating VaR.

Process:

1. Collect historical returns for the portfolio over a sufficiently long period (typically 250 to 500 trading days).
2. Apply each historical return to the current portfolio value to generate hypothetical profit and loss (P and L) figures.
3. Sort all hypothetical P and L outcomes from worst to best.
4. For a 95% one-day VaR using 250 days of data: the 5% worst-case threshold is the 12th or 13th worst observation (5% of 250 = 12.5).
5. That observation’s loss amount is the VaR.

Key characteristics:

• Does not assume normal distribution: captures actual fat tails and skewness in real data
• Easy to explain to non-technical stakeholders
• Depends heavily on the quality and relevance of the historical period chosen
• Does not capture risks that have never occurred in the historical sample
• Most commonly used by Indian asset managers and banks for practical VaR reporting

6. Method 3: Monte Carlo Simulation VaR

Monte Carlo simulation generates thousands or millions of hypothetical return scenarios using random sampling based on a defined statistical model of the portfolio’s return behaviour. VaR is then estimated from the resulting distribution of simulated outcomes.

The primary advantage of the Monte Carlo simulation is its ability to model complex financial instruments and portfolios with numerous risk factors. However, it is computationally intensive and requires the formulation of a mathematical model for asset returns, which may not always accurately represent reality.

Process:

1. Define a statistical model for the portfolio’s return drivers (correlations, volatility, distribution assumptions).
2. Run thousands of random simulations of possible portfolio outcomes over the chosen time horizon.
3. Sort the simulated outcomes from worst to best.
4. Identify the loss at the chosen confidence percentile. This is the Monte Carlo VaR.

Key characteristics:

• Most flexible: can model non-linear instruments, options, and complex portfolio structures
• Can incorporate user-defined scenarios including stressed correlations or fat-tail distributions
• Computationally demanding: requires significant computing resources
• Quality of output depends on the quality of the underlying model
• Used by sophisticated institutions including mutual funds with options strategies, hedge funds, and banks’ internal risk management teams

7. Worked Example 1: Parametric VaR for an Indian Portfolio

An investor holds a Rs 1 crore equity portfolio with an annual standard deviation of 20% (typical for a large-cap portfolio in India).

Required: One-day 95% VaR.

Step 1: Convert annual SD to daily SD

Daily SD = Annual SD / √252 (trading days in a year)
         = 20% / √252
         = 20% / 15.87
         = 1.26%

Step 2: Apply VaR formula at 95% confidence (Z = 1.65)

VaR = Portfolio Value x Daily SD x Z-score
    = Rs 1,00,00,000 x 0.0126 x 1.65
    = Rs 1,00,00,000 x 0.02079
    = Rs 2,07,900

Interpretation: With 95% confidence, this portfolio will not lose more than approximately Rs 2.08 lakh in a single trading day.

For a normal distribution, a 95% confidence level corresponds to 1.65 standard deviations. A portfolio of Rs 1 crore with 20% annual volatility would have a one-day 95% VaR of approximately Rs 2.08 lakhs. This means, with 95% confidence, the maximum loss in one day will not exceed Rs 2.08 lakhs.

8. Worked Example 2: Historical Simulation VaR

A portfolio manager has 250 daily return observations for a Rs 50 lakh debt fund. After sorting all 250 daily returns from worst to best, the observations at the 5% tail are as follows:

5% of 250 days = 12.5 observations. The 95% one-day historical VaR corresponds to the 12th or 13th worst return.

Taking the 12th worst: minus 1.4%

VaR = Rs 50,00,000 x 1.4% = Rs 70,000

Interpretation: Based on historical returns, there is a 5% probability that this portfolio could lose more than Rs 70,000 in a single day. This estimate does not assume a normal distribution and directly reflects the actual return behaviour of the portfolio.

9. Worked Example 3: Mutual Fund VaR Calculation

Using the parametric formula at a 95% confidence level for a mutual fund with a standard deviation of 2%, the VaR is approximately 2.068%. This means there is a 5% chance that on any given day, the fund could lose more than 2.068% of its value.

For a mutual fund investor with Rs 10 lakh invested:

One-day 95% VaR = Rs 10,00,000 x 2.068% = Rs 20,680

This investor can expect that on any given trading day, there is only a 5% probability of losing more than Rs 20,680. On 95 out of 100 trading days, the loss will be less than this amount.

10. Scaling VaR Across Time Horizons

VaR calculated for one time period can be scaled to another using the square root of time rule, provided returns are independent across periods:

VaR (T days) = VaR (1 day) x √T

Example: Convert one-day 99% VaR to 10-day 99% VaR:

10-day VaR = 1-day VaR x √10
           = Rs 2,07,900 x 3.162
           = Rs 6,57,382

Initial margin requirements at NSE are based on 99% value at risk over a one day time horizon. For futures contracts where it may not be possible to collect mark-to-market settlement before the commencement of trading on the next day, the initial margin is computed over a two-day time horizon by applying an appropriate statistical formula.

The square root of time rule holds only when daily returns are independent and identically distributed. In practice, volatility clustering (where volatile days tend to be followed by more volatile days) means this scaling can underestimate VaR during stressed market periods.

11. VaR in the Indian Regulatory Context: SEBI and NSE

VaR is not merely a theoretical concept in India. It is embedded in the regulatory and trading infrastructure of Indian capital markets.

As mandated by SEBI, the Value at Risk (VaR) margining system, which is internationally accepted as the best margining system, is applicable on the outstanding positions of the Members in all Securities on BSE. This system is used to contain the risk arising out of transactions entered into by members in various securities either on their own account or on behalf of their clients.

Daily margin payable by members comprises of the sum of VaR margin, Extreme Loss Margin, and mark-to-market margin.

SEBI’s VaR Margin Framework in Equity Cash Markets:

NSE and BSE compute and publish VaR margins for each stock on a daily basis. The VaR margin for a stock reflects the estimated maximum potential one-day loss at 99% confidence. Brokers are required to collect this margin upfront from clients before trades are executed.

For other securities categorised as Group III, VaR margin is applied at 8.66 times the Index VaR, reflecting the higher liquidity risk and volatility of less liquid securities.

This regulatory use of VaR directly affects every equity investor and trader in India. A stock with high volatility attracts a higher VaR margin, which limits the leverage that can be taken and protects the clearing system from defaults during market stress.

SEBI’s VaR Application to Liquid Funds:

NSE additionally publishes daily VaR haircut files for mutual fund units that are used as collateral for margin purposes. The provisions of the VaR margin rate framework for mutual fund units became effective from May 2, 2025, with members advised to refer to the MF haircut file available on the NSE website for daily updated figures.

12. Conditional VaR (CVaR): Beyond the Threshold

VaR tells you the threshold loss that will not be exceeded with a given probability. But it says nothing about what happens when that threshold is breached. In a severe market crash, VaR may indicate that the portfolio could lose Rs 2 lakh in a day at 95% confidence, but the actual loss in that 5% scenario could be Rs 5 lakh or Rs 10 lakh.

Conditional Value at Risk (CVaR), also called Expected Shortfall (ES) or Expected Tail Loss (ETL), addresses this gap.

CVaR is the expected loss given that the loss has already exceeded the VaR threshold. It is the average of all losses in the tail beyond the VaR cutoff.

CVaR = Expected value of all losses exceeding VaR

Maximum drawdown represents the peak-to-trough loss of an investment based on month-end closing values over a given period. This complements CVaR in capturing the severity of tail events that standard VaR does not address.

Example:

A portfolio has a one-day 95% VaR of Rs 2 lakh. Analysing the 5% worst outcomes from historical data:

• Day 1 of tail: loss of Rs 2.1 lakh
• Day 2 of tail: loss of Rs 2.5 lakh
• Day 3 of tail: loss of Rs 3.8 lakh

CVaR = Average of these tail losses = (2.1 + 2.5 + 3.8) / 3 = Rs 2.8 lakh

The CVaR of Rs 2.8 lakh tells the investor that if the 5% worst case materialises, they should expect to lose approximately Rs 2.8 lakh on average, not merely the Rs 2 lakh VaR threshold.

CVaR is mathematically more coherent than VaR because it satisfies all the properties of a coherent risk measure, particularly the property of sub-additivity (the risk of a combined portfolio is less than or equal to the sum of individual risks). Standard VaR can violate sub-additivity in certain scenarios. This makes CVaR increasingly preferred in regulatory and academic frameworks globally.

13. VaR vs Standard Deviation vs Downside Deviation

14. Advantages of VaR

Single, intuitive number: VaR converts complex portfolio risk into a single figure expressed in rupees or percentage. This makes it straightforward for non-technical stakeholders including boards, clients, and regulators to understand and act on.

VaR provides a single, summarising figure for the possible losses that can occur, making it simpler for investors and stakeholders to understand the risk level.

Versatility across asset classes: VaR is applicable to all types of assets: equities, bonds, derivatives, and more, making it a versatile risk assessment tool.

Regulatory acceptance: VaR is the globally recognised standard for financial institution risk measurement. It is the basis of Basel III capital adequacy requirements for banks and SEBI’s margining framework for Indian equity markets.

Comparable across portfolios: A VaR of Rs 50 lakh on a Rs 5 crore portfolio and a VaR of Rs 50 lakh on a Rs 50 crore portfolio carry very different messages about risk concentration. Expressing VaR as a percentage of portfolio value allows meaningful comparison.

Scenario analysis capability: VaR allows simulation of extreme market scenarios and helps understand how these conditions could affect the portfolio.

15. Limitations of VaR

VaR does not tell you what happens beyond the threshold: The most critical limitation. A 99% VaR of Rs 25 lakh says nothing about whether the 1% worst case will produce a loss of Rs 26 lakh or Rs 2.6 crore. CVaR addresses this gap.

Assumes normal market conditions: All three VaR methods are calibrated to normal market conditions. During extreme events like the COVID-19 crash of March 2020, actual losses far exceeded VaR estimates. VaR consistently underestimates tail risk during systemic crises.

Parametric VaR assumes normal distribution: Real return distributions have fat tails and negative skew. Assuming normality understates the probability of extreme losses.

Historical simulation is backward-looking: It assumes the future will resemble the past. A risk that has never occurred in the historical sample (a new type of crisis or a novel market structure failure) will not be captured.

VaR is not sub-additive: Standard VaR can produce counter-intuitive results where the VaR of a combined portfolio exceeds the sum of individual portfolio VaRs. CVaR does not have this problem.

Can create false confidence: A portfolio manager who shows a low VaR may mislead stakeholders into believing the portfolio is safe, even if large losses are possible in the tail scenarios that VaR does not describe.

16. VaR in Financial Planning Practice

For most retail financial planners in India, VaR is not calculated manually for client portfolios in daily practice. Its direct applications are more relevant to institutional contexts. However, understanding VaR is important for CFP professionals for three reasons.

First, SEBI’s VaR margin framework affects the leverage and margin requirements for all equity trading positions. A CFP advising a client who trades in futures and options must understand that the upfront margin requirement is directly derived from VaR.

Second, VaR is a statistical measure that indicates the maximum potential loss a portfolio could incur over a specified period, given a certain confidence level. Financial planners can use VaR alongside standard deviation, beta, and Sharpe ratio to build a complete risk picture for client portfolios.

Third, risk communication is a core financial planning skill. Being able to explain to a client that “your equity portfolio has a one-month 95% VaR of approximately Rs 3.5 lakh” is more actionable and specific than saying “your portfolio is risky.” It helps clients understand the practical implications of their asset allocation and encourages realistic expectations about short-term drawdown potential.

17. Comparison Table: Three VaR Methods

18. Key Exam Points

1. VaR Definition: Maximum expected loss over a specified period at a given confidence level under normal market conditions.
2. Three components of VaR: time horizon, confidence level, and loss amount. All three must be specified.
3. A one-day 95% VaR of Rs 1 lakh indicates there is only a 5% chance of losing more than Rs 1 lakh in a day under normal market conditions.
4. Three VaR methods: Parametric (assumes normal distribution, uses Z-score), Historical Simulation (uses actual past data, no distribution assumption), Monte Carlo (simulates thousands of scenarios, most flexible).
5. Z-scores for VaR: 90% confidence = 1.28; 95% confidence = 1.65; 99% confidence = 2.33.
6. Parametric VaR Formula: Portfolio Value x SD x Z-score x √Time.
7. Scaling across time horizons: VaR (T days) = VaR (1 day) x √T.
8. NSE initial margin requirements are based on 99% value at risk over a one-day time horizon. For futures contracts, the initial margin is computed over a two-day time horizon.
9. As mandated by SEBI, the VaR margining system is applicable on the outstanding positions of members in all securities on BSE, internationally accepted as the best margining system.
10. CVaR (Expected Shortfall): The average loss given that the loss has exceeded the VaR threshold. It answers “what do we expect to lose in the worst cases?” More informative than VaR for tail risk management.
11. VaR does not describe what happens beyond the threshold. CVaR addresses this limitation.
12. VaR assumes normal market conditions and can significantly underestimate risk during crises.

19. FAQs

What is Value at Risk (VaR) in simple terms? Value at Risk (VaR) is a statistical estimate of the maximum loss a portfolio is likely to suffer over a specific time period at a given level of confidence. For example, a one-day 95% VaR of Rs 2 lakh means there is a 5% probability of losing more than Rs 2 lakh in a single trading day. It combines three elements: a loss amount, a time period, and a probability level.

What are the three methods of calculating VaR? The three methods are: Parametric VaR (also called variance-covariance method, which assumes normal distribution and uses mean, standard deviation, and Z-score), Historical Simulation VaR (which ranks actual historical returns from worst to best and reads off the loss at the chosen percentile), and Monte Carlo VaR (which generates thousands of simulated portfolio scenarios using a defined statistical model).

How does SEBI use VaR in Indian markets? SEBI mandates VaR-based margin collection by brokers on all equity positions at NSE and BSE. NSE and BSE compute and publish daily VaR margin percentages for every listed security. These margins must be collected upfront from clients before trades. The initial margin for derivatives positions at NSE is based on 99% VaR over a one-day time horizon. SEBI also uses VaR haircuts for mutual fund units pledged as collateral.

What is the difference between VaR and CVaR? VaR defines a threshold: the maximum loss not exceeded at a given confidence level. CVaR (Conditional Value at Risk), also called Expected Shortfall, measures the average loss in the scenarios that exceed the VaR threshold. While VaR says “you will not lose more than Rs 2 lakh with 95% probability,” CVaR adds “and if you do exceed Rs 2 lakh, the expected loss is Rs 2.8 lakh.” CVaR is more informative for understanding tail risk.

What confidence level should be used for VaR in India? For regulatory purposes, SEBI and NSE use 99% confidence level for margin calculations on equity and derivative positions. For internal risk management and client reporting, 95% confidence level is more common as it offers a practical balance between sensitivity and accuracy. The choice depends on the purpose: more conservative institutional risk management uses 99%, while portfolio monitoring often uses 95%.

What are the main limitations of VaR? VaR’s key limitations are: it does not reveal what happens beyond the threshold (CVaR addresses this); it assumes normal market conditions and underestimates losses during crises; the parametric method assumes normally distributed returns, which underestimates fat tails; historical simulation depends on the past being representative of the future; and Monte Carlo simulation accuracy depends on the quality of the underlying statistical model.

20. CFP Exam Quick Recap

• VaR: Maximum loss not exceeded over a specified period at a given confidence level under normal conditions
• Three components: loss amount, time horizon, confidence level (all must be specified)
• Three methods: Parametric (Z-score, normal distribution); Historical (ranks actual past returns); Monte Carlo (simulated scenarios)
• Z-scores: 90% = 1.28; 95% = 1.65; 99% = 2.33
• Parametric VaR: Portfolio Value x SD x Z-score x √Time (daily SD = Annual SD / √252)
• Time scaling: VaR (T days) = VaR (1 day) x √T
• SEBI mandates 99% one-day VaR for NSE and BSE equity margin calculations
• NSE futures: initial margin based on 99% VaR over two-day time horizon
• CVaR (Expected Shortfall): Average loss beyond VaR threshold; more informative than VaR for tail risk
• VaR underestimates risk during crises; does not describe severity of tail losses; assumes normal conditions
• VaR is used in India for: equity margins (SEBI), mutual fund collateral haircuts, bank internal risk systems