Every capital budgeting decision involves future cash flows — and the future is uncertain. Risk Analysis in Capital Budgeting is simply a toolkit that helps you ask: what if things don't go as planned? The ICAI curriculum for Paper 6 expects you to know four main techniques, and this is frequently examined as an 8–10 mark problem.
Start with Sensitivity Analysis — the most exam-heavy technique. Here, you change one variable at a time (sales volume, selling price, cost of capital, etc.) while keeping everything else constant, and measure how much NPV or IRR changes. The output is usually expressed as a percentage change that makes NPV exactly zero. A variable with a small percentage tolerance (say 5%) is more sensitive — meaning your project is riskier to that variable. Think of it as finding your project's weakest link.
Next is Scenario Analysis, where instead of one variable, you change a bundle of variables together to create three full pictures: optimistic, pessimistic, and most likely. This gives you a range of NPVs — and from that range, you can judge overall risk. Unlike sensitivity analysis, it is more realistic because real life rarely changes just one thing at a time.
For adjusting the discount rate itself, we use the Risk-Adjusted Discount Rate (RADR) method. Here, riskier projects are evaluated at a higher discount rate — essentially demanding a higher return as compensation for risk. The logic: if a safe project needs 10% return, a risky one might need 15%. Simple, intuitive, widely used in practice.
The Certainty Equivalent (CE) method goes the other route — instead of adjusting the rate, you reduce the cash flows using a certainty equivalent coefficient (α, always between 0 and 1), then discount at the risk-free rate. If α = 0.8 on a ₹5,00,000 cash flow, you only count ₹4,00,000 as certain. Riskier years get lower α values.
Finally, Decision Tree Analysis is used for sequential decisions under uncertainty — drawn as a tree with decision nodes (squares) and chance nodes (circles), with probabilities and payoffs at each branch. You solve it by backward induction — rolling back from the end.