Value at Risk Explained for CAIIB Risk Management

Value at Risk is one of the most tested and most practically important concepts in the CAIIB Risk Management elective, because it puts a single rupee figure on how much a bank's portfolio could lose on a bad day. Whether you are revising market risk for the exam or trying to understand how your treasury desk sets limits, mastering Value at Risk gives you a clean mental model for measuring downside under uncertainty. This guide explains the concept, the calculation methods, confidence levels, regulatory context under Basel, and the limitations you must be able to critique in the exam.

What Value at Risk Means in Banking

Value at Risk (VaR) estimates the maximum loss a portfolio is expected to suffer over a defined holding period, at a chosen confidence level, under normal market conditions. It answers a precise question: "What is the worst loss I should expect, say, 99 times out of 100?"

A VaR statement always has three parts:

• A time horizon — commonly 1 day for trading books and 10 days for regulatory capital.
• A confidence level — typically 95% or 99%.
• A loss amount — expressed in currency.

For example: "The 1-day 99% VaR of the trading book is ₹5 crore." This means there is only a 1% chance the book loses more than ₹5 crore in a single day. Note that VaR tells you the threshold of loss, not the size of loss when that threshold is breached.

VaR is widely used for market risk limit-setting, capital allocation, and board-level risk reporting. Because it collapses many positions into one comparable number, it lets management compare risk across equity, forex, and fixed-income desks on a like-for-like basis. CAIIB candidates should be able to define VaR crisply and list its three components, as this forms the foundation for the rest of the market risk syllabus on the CAIIB course.

The Three Methods of Calculating Value at Risk

The CAIIB syllabus expects you to know all three standard approaches to computing Value at Risk, along with their assumptions and trade-offs.

1. Variance-Covariance (Parametric) Method

This assumes returns are normally distributed. VaR is derived from the portfolio's standard deviation (volatility) scaled by a confidence multiplier (z-score). The core formula is:

VaR = Z × σ × Portfolio Value × √t

where Z is the z-score (1.645 for 95%, 2.33 for 99%), σ is the daily volatility of returns, and √t scales a 1-day figure to a t-day horizon. It is fast but understates tail risk when returns are not normal.

2. Historical Simulation

This re-prices the current portfolio using actual historical return data (e.g., the last 250 trading days), sorts the simulated profit-and-loss outcomes, and reads off the loss at the chosen percentile. It makes no distributional assumption but assumes the past represents the future.

3. Monte Carlo Simulation

This generates thousands of random return scenarios from a chosen statistical model, re-values the portfolio in each, and takes the percentile loss. It is the most flexible and handles non-linear instruments like options well, but it is computationally heavy. Practise these contrasts on the CAIIB mock tests to lock in the differences.

Confidence Levels and Holding Periods

Getting the Value at Risk parameters right is where many candidates lose marks. The confidence level sets how conservative the estimate is, and the holding period reflects how long it would take to liquidate or hedge positions.

Standard Confidence Multipliers

• 95% confidence → z-score of 1.645
• 99% confidence → z-score of 2.33
• 97.5% confidence → z-score of 1.96 (used in the Expected Shortfall framework)

A higher confidence level produces a larger VaR number, because you are estimating a rarer, deeper loss. Internal trading desks often use a 1-day 95% or 99% VaR for daily limit monitoring, while regulatory capital traditionally used a 10-day 99% VaR.

Scaling Across Time

Under the square-root-of-time rule, a 1-day VaR is converted to an n-day VaR by multiplying by √n. So a 10-day VaR is roughly the 1-day VaR × √10 (about 3.16×). This shortcut assumes returns are independent and identically distributed day to day — an assumption that breaks down during volatility clustering and crises. Keep an eye on prevailing benchmark yields and policy moves on the RBI rates tracker, since shifts in rates feed directly into the volatility inputs your VaR model uses.

Value at Risk, Expected Shortfall and Basel

The biggest weakness of Value at Risk is that it says nothing about losses beyond the threshold. A 99% VaR ignores the severity of the worst 1% of outcomes. To fix this, regulators and risk managers increasingly rely on Expected Shortfall (ES), also called Conditional VaR.

Expected Shortfall

ES is the average loss given that the loss has exceeded the VaR threshold. It captures tail severity and is "coherent" (it respects diversification, which VaR does not always do). Under the Basel Fundamental Review of the Trading Book (FRTB), the internal models approach replaced 99% VaR with a 97.5% Expected Shortfall measure for market risk capital, calibrated to a stressed period.

Backtesting Requirement

Basel requires banks using internal models to backtest VaR: compare daily VaR estimates against actual P&L. Too many "exceptions" (days where loss exceeds VaR) push the bank into Basel's penalty zones:

• Green zone — up to 4 exceptions in 250 days (acceptable).
• Yellow zone — 5 to 9 exceptions (a rising capital multiplier applies).
• Red zone — 10 or more exceptions (model rejected).

Understanding how VaR plugs into market risk capital is exactly the kind of integrated question CAIIB favours. Follow regulatory updates on the IIBF news page so your exam answers reflect current Basel norms.

Limitations You Must Be Able to Critique

Examiners reward candidates who can criticise Value at Risk, not just compute it. The key limitations are:

• It ignores tail magnitude — VaR gives the threshold, not the expected size of catastrophic losses (which is why ES exists).
• Normality assumption — the parametric method assumes a bell curve, but real returns have "fat tails" and skewness, so extreme events occur more often than predicted.
• Backward-looking data — historical simulation assumes the recent past repeats; it misses structural breaks and regime changes.
• Non-subadditivity — in some cases the VaR of a combined portfolio can exceed the sum of individual VaRs, violating the diversification principle.
• Procyclicality — VaR shrinks in calm markets (encouraging risk-taking) and spikes in crises (forcing fire-sales).

This is why VaR is always paired with stress testing and scenario analysis, which probe extreme but plausible events that history may not contain. A complete market risk framework uses VaR for day-to-day limits, ES for capital, and stress tests for the unexpected. Reinforce these critiques with quick recall drills on match-the-concept games and browse related explainers on the iibf.store blog.

For authoritative guidance, refer to the official resources of the Reserve Bank of India and the Indian Institute of Banking & Finance.

Frequently Asked Questions

Conclusion: Turn Theory into Marks

Mastering Value at Risk — its definition, three calculation methods, confidence multipliers, the move to Expected Shortfall under Basel, and its limitations — covers a large slice of the CAIIB market risk syllabus. The fastest way to retain it is active practice. Attempt timed questions on the CAIIB Risk Management mock tests and structure your full preparation through the CAIIB course on iibf.store to walk into the exam confident on every VaR question.