Value at Risk in Treasury Portfolios: A TIRM Guide (2026)

Every treasury desk needs a single number that tells top management "how much can we lose tomorrow?" That number is Value at Risk (VaR). For TIRM candidates, value at risk in treasury portfolios is one of the most heavily tested market-risk concepts because it links statistics, regulatory capital, and everyday desk decisions into one framework. This guide breaks down what VaR measures, how banks calculate it for bond and money-market holdings, where it fits against other risk yardsticks, and the RBI/Basel expectations around it — all pitched at exam difficulty.

📊 What Is Value at Risk (VaR) in Treasury Risk Management?

Value at Risk answers a precise question: over a given holding period and at a given confidence level, what is the maximum loss a portfolio is expected NOT to exceed under normal market conditions? A "1-day 99% VaR of ₹5 crore" means that, on 99 days out of 100, the treasury book should not lose more than ₹5 crore in a single trading day. The remaining 1% is the "tail" — the rare, larger loss that VaR deliberately does not size, which is why banks pair it with stress testing.

VaR became central to treasury oversight because a single duration or PV01 figure tells you sensitivity to a rate move, but not the probability-weighted rupee loss across an entire portfolio of G-Secs, SDLs, corporate bonds, and money-market paper. VaR converts that mixed basket into one comparable number that the ALCO and the board can monitor daily. Understanding this framework is easier once you have worked through the fundamentals in the Risk Analysis and Control chapter, which lays out the full toolkit of market risk metrics before VaR is layered on top.

📌 Remember: VaR is a statistical estimate of "normal" loss, not a worst-case number — stress tests and scenario analysis exist precisely to cover what VaR excludes.

🧮 How Banks Compute VaR for Investment Portfolios

Three methods dominate exam questions and real desks alike. Historical simulation re-prices today's portfolio using actual daily price/yield changes from a past window (commonly 250 trading days), ranks the resulting P&L outcomes, and reads off the loss at the chosen percentile. It needs no assumption about the shape of returns but is only as good as the historical window chosen. Variance-covariance (parametric) VaR assumes returns are normally distributed and derives VaR from the portfolio's standard deviation and a confidence multiplier (1.65 for 95%, 2.33 for 99%); it is fast to compute but understates risk when markets are fat-tailed, exactly the periods that matter most. Monte Carlo simulation generates thousands of random price paths from an assumed distribution and rebuilds the full loss distribution — the most flexible and computationally heaviest of the three.

Banks scale VaR by the square-root-of-time rule to move between horizons (a 10-day VaR is roughly the 1-day VaR multiplied by √10) and validate their models through backtesting — comparing predicted VaR breaches against actual daily losses. Too many breaches versus the confidence level signals a model that is understating risk. This computational discipline builds directly on the sensitivity measures covered under Bond duration and PV01, since duration and PV01 feed into the price-shock inputs used in parametric VaR.

💡 Exam Tip: If a question asks which VaR method is fastest but weakest in tail-risk periods, the answer is the variance-covariance (parametric) method.

⚖️ VaR vs Other Market Risk Measures

TIRM papers frequently test whether candidates can distinguish VaR from adjacent metrics rather than compute it in isolation. The table below is the quick-reference version worth memorising before the exam.

Notice that VaR and duration/PV01 are complementary, not competing — duration tells you sensitivity per basis point, VaR translates that sensitivity into a rupee-loss probability across the whole book. Expected Shortfall (Conditional VaR) is increasingly favoured by regulators globally because, unlike plain VaR, it does look inside the tail and averages the losses beyond the cutoff, which is why Basel's market risk framework has been shifting capital calculations toward it. Candidates preparing money-market topics should also revisit money market instruments, since short-tenor paper has its own liquidity-risk overlay on top of price VaR.

🏦 RBI Guidelines and Basel Norms on VaR-Based Capital

Indian banks compute capital charges for market risk in the trading book broadly along Basel lines, which permit internal VaR-based models (subject to regulatory approval) as an alternative to the standardised measurement method. RBI's guidelines on classification and valuation of investments prescribe the risk-management architecture — including independent risk measurement, backtesting, and board-level oversight — within which VaR models must operate; treasury desks are expected to reconcile VaR-based limits with mark-to-market outcomes reported for AFS/HFT books. For the authoritative and most current text, refer directly to the RBI's Master Direction on investment classification and valuation rather than relying on secondary summaries, since limits and disclosure formats are periodically revised.

Internally, banks set VaR-based intraday and overnight limits for each desk, aggregate them at the treasury level, and escalate breaches to the ALCO. This ties back to the governance structure taught in the Risk Management Process chapter, where limit-setting, monitoring, and escalation form one continuous cycle rather than isolated steps.

⚠️ Common Mistake: Candidates often assume VaR alone satisfies regulatory capital requirements — in practice it must sit alongside stress testing, backtesting, and board-approved limit structures.

🎯 Common Pitfalls in Applying VaR

Three mistakes recur both in the exam hall and on real desks. First, treating VaR as a worst-case figure — it is explicitly a "normal market conditions" estimate, and every VaR disclosure should be read alongside a stress-VaR or scenario number. Second, ignoring model risk: parametric VaR under-predicts losses when correlations between instruments break down in a crisis, exactly when accurate numbers matter most — 2008 and the 2013 taper-tantrum bond sell-off are the standard illustrative episodes. Third, using a stale historical window; a 250-day lookback computed entirely during a calm period will produce an artificially low VaR the moment volatility returns.

Good treasury practice therefore layers VaR with duration/PV01 limits, stress tests, and concentration limits rather than relying on any single metric — much like the discipline demanded by mark to market valuation of bank investments, where a single day's price movement must be reconciled across the whole AFS/HFT book, not just flagged desk-by-desk. This layered approach is exactly what separates a book that survives a rate shock from one that discovers its risk model too late.

🧠 Practice MCQs: Value at Risk in Treasury Portfolios

Want chapter-wise mock tests with 100+ MCQs? Start practising free →

❓ Frequently Asked Questions

VaR is one number among several a treasury desk must watch — pair it with duration, PV01, and mark-to-market discipline for the full picture, then test your understanding with our TIRM topic hub and a full practice test before exam day.