Sensitivity, Volatility and Value-at-Risk Interpretation - Article

Introduction

Market risk arises when the value of a bank's financial position changes because market variables move. Important variables include interest rates, foreign exchange rates, equity prices, commodity prices and market volatility.

A bank therefore needs tools that answer three different questions:

• Sensitivity: How much may the value change when a particular market factor changes?

• Volatility: How much does that market factor normally fluctuate?

• Value-at-Risk (VaR): How large could the portfolio loss be over a specified period at a stated confidence level?

These concepts are closely connected but are not interchangeable.

For CAIIB, the important skill is not merely remembering definitions. A learner should be able to interpret a risk measure correctly, compare different exposures and identify misleading statements.

The current Basel market-risk framework also treats sensitivity as an important risk concept. It defines sensitivity as the estimated change in the value of an instrument due to a small change in an underlying risk factor. Delta and vega are examples of sensitivities.

Learning Objectives

After studying this topic, you should be able to:

• explain sensitivity in market-risk measurement;

• understand the meaning of market volatility;

• distinguish sensitivity from volatility;

• interpret VaR correctly using amount, time horizon and confidence level;

• understand the main factors that influence VaR;

• recognise the limitations of VaR; and

• avoid common CAIIB traps involving these risk measures.

Core Relationship: Exposure, Sensitivity and Market Movement

A financial instrument does not lose money merely because a market variable exists. Loss occurs when the bank has an exposure whose value responds to movement in that variable.

The basic relationship can be understood in three stages:

Market risk factor → Movement in risk factor → Change in position value

Suppose a bank holds a fixed-rate bond.

Its market value is exposed to interest-rate movements. If market yields rise, the bond price generally falls. The amount by which the bond price responds depends on its interest-rate sensitivity.

Now suppose interest rates are highly volatile. Large changes in yield become more likely. A highly sensitive bond exposed to a highly volatile interest-rate environment may therefore have substantial market risk.

VaR goes one step further. It combines information about exposure, market movements and portfolio behaviour to express risk as a potential monetary loss over a specified horizon and confidence level.

Understanding Sensitivity

What Is Sensitivity?

Sensitivity measures how much the value of a financial instrument or portfolio is expected to change when a relevant market risk factor changes by a small amount.

The Basel market-risk terminology similarly describes sensitivity as a bank's estimate of the change in an instrument's value resulting from a small change in an underlying risk factor.

The risk factor may be:

• an interest rate;

• a bond yield;

• an exchange rate;

• an equity price;

• a commodity price;

• a credit spread; or

• volatility itself.

Sensitivity therefore links the market variable with the value of the exposure.

Positive and Negative Sensitivity

The direction of sensitivity is also important.

A positive sensitivity generally means that the value of the position increases when the relevant risk factor increases.

A negative sensitivity means that the position value decreases when that risk factor increases.

The exact interpretation depends on the instrument and the way the risk factor is defined.

For example, the price of a conventional fixed-rate bond normally moves in the opposite direction to its yield. Therefore, an increase in yield generally reduces its market value.

Higher Sensitivity Means Greater Response

Consider two bonds exposed to the same change in interest rates.

If Bond A is more interest-rate sensitive than Bond B, the same yield movement will normally produce a larger price movement in Bond A.

This does not automatically mean Bond A will always suffer a larger loss. The direction of the market movement and position also matters.

Common Sensitivity Measures

Banks use different sensitivity measures for different instruments.

Under the current Basel standardised market-risk approach, sensitivities are used for delta, vega and curvature risk calculations across prescribed risk classes.

For CAIIB, the key idea is that sensitivity is a response measure, not a probability measure.

Understanding Volatility

Meaning of Volatility

Volatility describes the degree to which a market price, rate or return fluctuates over time.

Higher volatility means larger or more frequent movements.

Lower volatility means movements are generally more stable.

For example:

• a currency moving within a narrow range has relatively low volatility;

• a currency showing sharp daily movements has higher volatility.

Volatility therefore represents uncertainty or variability in market movements.

Sensitivity and Volatility Are Different

This distinction is extremely important.

Sensitivity belongs primarily to the position.

Volatility belongs primarily to the market risk factor or its return behaviour.

A position may be highly sensitive even when market volatility is temporarily low. Similarly, the market may become highly volatile while a particular position has very little exposure to that market factor.

Why Volatility Matters for Options

Volatility is particularly important for options.

An option's value depends partly on expected future movement in the underlying price. Other factors remaining unchanged, greater expected volatility generally increases the value of optionality because there is greater scope for favourable price movement.

Sensitivity to changes in implied volatility is called vega risk. Basel defines vega risk as the potential loss arising from a change in the value of a derivative due to a change in the implied volatility of its underlying.

What Is Value-at-Risk?

Value-at-Risk is a statistical measure used to express market risk as a potential loss over a specified time horizon and at a specified confidence level.

Basel describes VaR as a measure of the worst expected portfolio loss resulting from market movements over a given time horizon and at a predefined confidence level.

Three elements must therefore always accompany a VaR figure:

1. Loss amount

2. Time horizon

3. Confidence level

Without these three elements, a VaR number cannot be interpreted properly.

Correct Interpretation of VaR

Suppose a bank reports:

One-day 99% VaR = ₹10 crore

The correct interpretation is:

Under the assumptions and data used by the VaR model, the bank expects the loss to remain within ₹10 crore on approximately 99% of comparable one-day periods.

Correspondingly, there remains approximately a 1% tail probability that the loss may exceed ₹10 crore.

This interpretation contains several important points.

VaR Is Not the Maximum Possible Loss

A 99% VaR of ₹10 crore does not mean that the bank cannot lose more than ₹10 crore.

Losses beyond the VaR threshold are possible.

VaR mainly identifies a percentile or threshold in the loss distribution. It does not by itself tell the bank how severe a loss could become once that threshold is crossed.

VaR Does Not Predict the Exact Next-Day Loss

VaR is a risk estimate, not a forecast that tomorrow's loss will equal the VaR figure.

The actual result may be:

• a profit;

• a small loss;

• a loss near the VaR amount; or

• an extreme loss greater than VaR.

Confidence Level Matters

A VaR figure is meaningless without its confidence level.

A 99% VaR will normally be higher than a 95% VaR for the same portfolio, data and time horizon because the 99% measure looks further into the tail of the loss distribution.

Therefore:

Higher confidence level → generally higher VaR

Time Horizon Matters

A one-day VaR and a ten-day VaR cannot be directly treated as the same measure.

A longer holding period allows more time for adverse market movements to accumulate.

Therefore, other assumptions remaining comparable:

Longer risk horizon → generally greater potential loss

However, the relationship need not be perfectly linear.

Factors That Influence VaR

Several factors can change a portfolio's VaR.

1. Size of Position

Larger market exposures generally create larger potential gains or losses.

Other factors remaining unchanged, increasing the size of an exposure usually increases VaR.

2. Sensitivity

A position that reacts strongly to market-factor movements can produce larger changes in value.

Higher sensitivity can therefore increase portfolio risk.

3. Market Volatility

When market volatility rises, larger price or rate movements become more likely.

Other factors remaining unchanged:

Higher volatility generally leads to higher VaR.

4. Confidence Level

Moving from a lower confidence level to a higher confidence level usually increases the loss threshold represented by VaR.

5. Time Horizon

A longer holding or risk period generally results in a higher potential loss estimate.

6. Correlation

Portfolio positions do not always move together.

If different positions are imperfectly correlated, losses in one position may be partly offset by gains or smaller losses in another.

This is the basis of the diversification effect.

Basel defines diversification as the reduction in portfolio risk arising from positions that are not perfectly correlated.

However, correlations can change sharply during stressed markets. Diversification observed in normal conditions may therefore weaken during a crisis.

Sensitivity, Volatility and VaR Compared

VaR Methodology: What the Learner Should Know

There are three commonly discussed approaches to estimating VaR.

Historical Simulation

Historical market movements are applied to the current portfolio to estimate how it would have performed under past market conditions.

Its major strength is that it does not require a normal-distribution assumption in the same way as a simple parametric model.

Its major weakness is dependence on the selected historical period. Risks not represented in that data may be underestimated.

Variance-Covariance or Parametric Approach

This method uses statistical estimates such as volatility and correlation, together with assumptions regarding return behaviour.

It is comparatively efficient for portfolios that behave approximately linearly.

However, simplifying distribution assumptions may become unreliable during extreme markets or for portfolios containing significant optionality.

Monte Carlo Simulation

A large number of possible market scenarios are generated using a statistical model and the portfolio is revalued under those scenarios.

It can handle complex portfolios more flexibly but requires greater modelling effort, computing resources and valid assumptions.

For this concept-based topic, the important examination point is that different VaR methodologies can produce different VaR numbers for the same portfolio because their assumptions and data treatment differ.

Why VaR Is Useful to a Bank

VaR provides a common monetary language for different kinds of market risk.

It can support:

• treasury risk monitoring;

• comparison of portfolios;

• internal risk limits;

• allocation of risk capacity;

• management reporting;

• assessment of diversification;

• identification of unusual changes in exposure; and

• discussion between dealers, risk managers and senior management.

A treasury portfolio may simultaneously contain bonds, foreign-exchange positions and derivatives. Individual sensitivity measures remain important, but VaR can provide a consolidated view of market-risk exposure.

VaR should therefore complement—not replace—position limits, sensitivity limits, stop-loss limits, stress tests and other controls.

Limitations of VaR

VaR Does Not Measure Tail Severity

Knowing that losses may exceed VaR does not reveal how large those losses might become.

This is one reason tail-risk measures and stress testing are important.

Basel defines Expected Shortfall (ES) as the average of potential losses exceeding VaR at a given confidence level.

Historical Data May Not Represent the Future

VaR estimates often depend heavily on historical prices, volatilities and correlations.

A future crisis may behave differently from the historical observation period.

Correlations Can Change

Assets that appear diversified during normal periods may start moving in the same adverse direction during market stress.

Model Risk Exists

Different assumptions, data windows and valuation techniques may produce materially different results.

A mathematically precise VaR figure can still be misleading if the model is inappropriate.

Non-linear Positions Are More Difficult

Options and other instruments with non-linear payoffs may not be adequately represented by simple linear sensitivity assumptions.

VaR Can Create False Comfort

A low VaR number should never be interpreted as proof that a portfolio is safe.

Concentration, liquidity risk, event risk and extreme market movements may remain significant.

VaR and the Current Basel Perspective

For exam purposes, VaR remains an important market-risk concept and its interpretation, assumptions, methodologies, limitations and backtesting remain relevant in IIBF risk-management syllabus material.

At the same time, learners should not assume that VaR is the only modern regulatory market-risk measure.

Under the current Basel market-risk framework:

• the standardised approach uses risk sensitivities including delta, vega and curvature; and

• the internal models framework uses Expected Shortfall as an important regulatory risk measure, while VaR continues to have a role in model backtesting.

This distinction is useful because an exam statement such as “VaR has completely disappeared from modern market-risk management” would be misleading.

Practical Banking Application

Consider a bank treasury holding government securities and foreign-exchange positions.

The risk-management function may examine the portfolio at several levels:

• Position amount tells how much exposure exists.

• Duration or PVBP indicates interest-rate sensitivity.

• FX sensitivity shows the effect of exchange-rate movements.

• Volatility indicates how unstable the relevant market factors are.

• VaR provides a consolidated estimate of potential loss under the selected model.

• Stress tests examine severe but plausible adverse situations.

• Limits and stop-loss controls restrict risk-taking.

No single measure provides the complete picture.

A sound market-risk framework therefore uses several complementary measures.

CAIIB Exam Focus

Remember these high-value concepts:

• Sensitivity measures the response of a position to a market-factor movement.

• Volatility measures the variability of the market factor or return.

• High sensitivity and high volatility together can create substantial market risk.

• VaR must always specify loss amount + time horizon + confidence level.

• VaR is not maximum possible loss.

• A 99% VaR still leaves a tail in which losses may exceed the VaR amount.

• Higher volatility generally increases VaR, other factors remaining unchanged.

• Higher confidence level generally increases VaR.

• Longer time horizon generally increases the potential-loss estimate.

• Diversification can reduce portfolio risk when positions are not perfectly correlated.

• Historical relationships and correlations may fail during stressed markets.

• Different VaR methodologies may produce different results.

• VaR should be supplemented by stress testing.

• Expected Shortfall focuses on losses beyond the VaR threshold.

• Delta measures linear sensitivity, while vega captures sensitivity to implied volatility.

Common Exam Traps

Trap 1: “VaR represents the maximum possible loss.”

Incorrect. Losses greater than VaR can occur. VaR identifies a loss threshold associated with a confidence level.

Trap 2: “A 99% VaR means there is a 99% probability that the portfolio will make a profit.”

Incorrect. The confidence level relates to the specified loss threshold, not to the probability of earning a profit.

Trap 3: “Higher volatility always means the position itself is more sensitive.”

Incorrect. Volatility describes market movement; sensitivity describes the position's reaction to that movement.

Trap 4: “A higher confidence level produces a lower VaR.”

Normally incorrect. Other factors remaining unchanged, a higher confidence level usually produces a higher VaR because the measure moves further into the loss tail.

Trap 5: “Diversification eliminates market risk.”

Incorrect. Diversification may reduce portfolio risk but cannot guarantee elimination of losses. Correlations may also increase during stressed conditions.

Trap 6: “VaR removes the need for stress testing.”

Incorrect. VaR may not adequately capture extreme events. Stress testing examines severe conditions outside normal model assumptions.

Trap 7: “Two portfolios with the same VaR must have identical risk.”

Incorrect. They may have different concentrations, sensitivities, liquidity characteristics and tail-loss profiles despite reporting the same VaR.

Trap 8: “Expected Shortfall and VaR measure exactly the same thing.”

Incorrect. VaR identifies a loss threshold at a confidence level. Expected Shortfall focuses on the average loss beyond that threshold.

Memory Technique

Use S–V–V:

S — Sensitivity: How strongly does my position react?V — Volatility: How much can the market move?V — VaR: What loss threshold does the combined risk indicate?

For interpreting VaR, remember A–T–C:

A — AmountT — Time horizonC — Confidence level

A VaR figure without all three is incomplete.