Strategic Financial Management · Investment Decisions, Project Planning and Control

Risk Analysis in Capital Budgeting: Methods and Numericals

Updated 11 October 2026 · Fact-checked

Risk analysis in capital budgeting tests how safe a project's NPV is when cash flows are uncertain. You either adjust the discount rate (RADR), shrink the cash flows (certainty equivalent), or test the NPV under changed inputs (sensitivity, scenarios, simulation, decision trees, expected NPV with standard deviation). Then you recommend accept or reject.

Understand Risk Analysis in Capital Budgeting

A project's cash flows are forecasts, not facts. Risk here means the cash flows can differ from the forecast, and you can attach probabilities to the outcomes. Plain NPV uses one best estimate and hides this. Risk analysis shows how much the answer can move.

There are two ways to build risk into the NPV itself. The risk-adjusted discount rate (RADR) keeps cash flows as forecast but raises the discount rate above the risk-free rate for riskier projects. The certainty equivalent (CE) method cuts each cash flow to the sure amount you would accept instead, using a coefficient between 0 and 1, and then discounts at the risk-free rate. RADR puts all risk in the rate. CE puts it in the cash flows, so it can treat each year's risk separately.

The other tools test the NPV instead of changing the rate. Sensitivity analysis changes one input at a time (sales volume, price, cost) and shows how far NPV moves. Scenario analysis changes several inputs together into pessimistic, most likely and optimistic cases. Simulation (Monte Carlo) draws random values for the uncertain inputs many times and builds a distribution of NPV.

When you have probabilities, work out the expected NPV and the standard deviation of NPV. A higher standard deviation means more risk. A decision tree is used when decisions come in stages: each branch has a probability and a cash flow, and you roll the tree back from the end to find the best first decision.

Key rules to remember

Risk-adjusted discount rate: RADR = Risk-free rate + Risk premium; NPV = Σ [CFt ÷ (1 + RADR)^t] − Initial outlay; Use the higher rate for riskier projects. Cash flows are the expected (uncertain) ones.
Certainty equivalent coefficient: α(t) = Certain cash flow ÷ Risky expected cash flow; α lies between 0 and 1. It usually falls for later years as uncertainty grows.
NPV by certainty equivalent: NPV = Σ [α(t) × CFt ÷ (1 + Rf)^t] − Initial outlay; Discount at the risk-free rate only. Using RADR here counts risk twice.
Expected NPV: E(NPV) = Σ P(i) × NPV(i); P(i) are the probabilities of the scenarios and they must add up to 1.
Standard deviation of NPV (scenarios): σ = √[ Σ P(i) × (NPV(i) − E(NPV))² ]; Use the NPV of each scenario, not the cash flow.
Standard deviation of NPV (yearly cash flows, independent): σ(NPV) = √[ Σ σt² ÷ (1 + r)^(2t) ]; Valid when cash flows of different years are independent. Use the risk-free rate unless the question says otherwise.
Standard deviation of NPV (perfectly correlated): σ(NPV) = Σ [ σt ÷ (1 + r)^t ]; Valid only when cash flows of all years are perfectly correlated.
Coefficient of variation: CV = σ ÷ E(NPV); Risk per rupee of expected NPV. Compare projects with different sizes. Lower is better. Meaningful only when E(NPV) is positive.
Probability of NPV below zero (normal assumption): Z = (0 − E(NPV)) ÷ σ; read the area from the normal table; Use only when the question tells you to assume NPV is normally distributed.
Sensitivity of a variable: % change that makes NPV zero = NPV ÷ PV of that variable's cash flows × 100; A smaller percentage means NPV is more sensitive to that variable.
Decision tree roll-back: EMV at a chance node = Σ (probability × value of branch); at a decision node choose the highest EMV; Work from right to left. Deduct outlays on the branch where they occur.

How to solve Risk Analysis in Capital Budgeting questions

Read the question for the method it names or the data it gives. Then follow this order so that no marks are lost.

1Identify the method from the data. Risk-free rate with coefficients means CE. A risk premium or a higher rate means RADR. Probabilities with scenarios mean expected NPV and standard deviation. Stages with branches mean a decision tree.
2List the initial outlay and the yearly cash flows. Check whether the cash flows are after tax and whether depreciation has been handled.
3Choose the right rate: RADR for risky cash flows, risk-free rate for CE flows, cost of capital for scenario NPVs unless told otherwise.
4Compute the present value factors to four decimals and build a small table: year, cash flow (or CE flow), factor, present value.
5Find the NPV. For probabilistic data, find the NPV of each scenario first, then the expected NPV, then the deviations, then the standard deviation.
6For sensitivity, change one variable at a time, recompute NPV, and rank the variables by how much NPV moves.
7For a decision tree, draw the tree, mark decision squares and chance circles, then roll back from the right.
8Write a clear recommendation: accept if NPV is positive and the risk is acceptable. Mention the standard deviation or CV if you have it.

Quickest way: Shortcut for RADR vs CE and expected NPV questions

When to use it: Use under time pressure when the question gives a few years of cash flows and one or two scenarios.

Write the PV factors once and reuse them. For equal annual flows, use the annuity factor.
For CE, multiply each cash flow by α first, then by the risk-free factor. Do not touch RADR.
For scenarios with equal annual cash flows, compute NPV = annual flow × annuity factor − outlay for each scenario. Only the annual flow changes.
Work in ₹ thousand for the standard deviation to avoid very large squares, and convert back at the end.
If E(NPV) is negative, you can state reject straight away. Still show the standard deviation if asked.

Common mistakes in Risk Analysis in Capital Budgeting

Discounting certainty equivalent cash flows at the risky rate (RADR).
Students use the cost of capital out of habit.
Fix: CE flows already carry the risk adjustment. Discount them at the risk-free rate.
Taking the average of the cash flows as the answer to a probability question and skipping NPV of each scenario.
It looks quicker, and students forget that the standard deviation needs NPV deviations.
Fix: Compute NPV for each scenario, then the expected NPV, then the standard deviation of NPV.
Changing several variables together in a sensitivity analysis.
Students mix up sensitivity and scenario analysis.
Fix: Sensitivity changes one variable at a time with the rest held constant. Scenario analysis changes a set of variables together.
Rolling back a decision tree from the left, or ignoring the cost at a later decision node.
Students read the tree like a normal flow chart.
Fix: Start at the end branches, calculate EMV at each chance node, pick the best option at each decision node, deduct the outlay on its own branch, then move left.
Using the independent-flows standard deviation formula when the question says cash flows are perfectly correlated, or the reverse.
Students memorise one formula only.
Fix: Read the correlation statement. Independent: add discounted variances and take the root. Perfectly correlated: add discounted standard deviations.
Writing only the numbers and no recommendation.
Students feel the calculation is the answer.
Fix: Finish with accept or reject and give the reason: expected NPV, risk measure and, if relevant, the more sensitive variable.

Worked examples

Example 1

Prime Auto Components Ltd is evaluating a project costing ₹10,00,000. Expected cash flows are ₹4,00,000, ₹5,00,000 and ₹6,00,000 in years 1 to 3. The risk-free rate is 6% and the risk-adjusted discount rate is 12%. Certainty equivalent coefficients for years 1 to 3 are 0.90, 0.80 and 0.70. Find the NPV by (a) RADR method and (b) CE method and advise. PV factors at 12%: 0.8929, 0.7972, 0.7118. PV factors at 6%: 0.9434, 0.8900, 0.8396.

Show the solution

(a) RADR method. Year 1: 4,00,000 × 0.8929 = ₹3,57,160.
Year 2: 5,00,000 × 0.7972 = ₹3,98,600.
Year 3: 6,00,000 × 0.7118 = ₹4,27,080.
Total PV = 3,57,160 + 3,98,600 + 4,27,080 = ₹11,82,840. NPV = 11,82,840 − 10,00,000 = ₹1,82,840.
(b) CE method. Certain cash flows: 4,00,000 × 0.90 = ₹3,60,000; 5,00,000 × 0.80 = ₹4,00,000; 6,00,000 × 0.70 = ₹4,20,000.
Discount at 6%: 3,60,000 × 0.9434 = ₹3,39,624; 4,00,000 × 0.8900 = ₹3,56,000; 4,20,000 × 0.8396 = ₹3,52,632.
Total PV = 3,39,624 + 3,56,000 + 3,52,632 = ₹10,48,256. NPV = 10,48,256 − 10,00,000 = ₹48,256.

Answer: NPV is ₹1,82,840 by RADR and ₹48,256 by CE. Both are positive, so the project should be accepted. The CE method gives a lower NPV because it reduces the later cash flows more heavily. The project passes under the stricter test as well.

Example 2

Shakti Textiles Ltd is considering a machine costing ₹10,00,000 with a 3-year life and no salvage value. Annual after-tax cash flow depends on demand: pessimistic (probability 0.3) ₹3,00,000; most likely (0.5) ₹4,00,000; optimistic (0.2) ₹5,00,000. The cost of capital is 10%. The 3-year annuity factor at 10% is 2.4869. Calculate the NPV in each case, the expected NPV and the standard deviation of NPV, and advise.

Show the solution

Pessimistic NPV = 3,00,000 × 2.4869 − 10,00,000 = 7,46,070 − 10,00,000 = −₹2,53,930.
Most likely NPV = 4,00,000 × 2.4869 − 10,00,000 = 9,94,760 − 10,00,000 = −₹5,240.
Optimistic NPV = 5,00,000 × 2.4869 − 10,00,000 = 12,43,450 − 10,00,000 = ₹2,43,450.
Expected NPV = 0.3 × (−2,53,930) + 0.5 × (−5,240) + 0.2 × 2,43,450 = −76,179 − 2,620 + 48,690 = −₹30,109.
Deviations from expected NPV (₹ thousand): pessimistic −253.930 − (−30.109) = −223.821; most likely −5.240 + 30.109 = 24.869; optimistic 243.450 + 30.109 = 273.559.
Squares (₹ thousand squared), approximately: 50,095.84; 618.47; 74,834.53.
Variance = 0.3 × 50,095.84 + 0.5 × 618.47 + 0.2 × 74,834.53 = 15,028.75 + 309.23 + 14,966.91 = 30,304.89.
Standard deviation = √30,304.89 ≈ 174.08 thousand, that is about ₹1,74,083.

Answer: Expected NPV is −₹30,109 and the standard deviation is about ₹1,74,083. The expected NPV is negative, and the spread is large relative to it. Reject the machine unless there are strategic benefits not captured in the cash flows.

Exam tips

Match the method to the data: Rf and α mean CE, a premium or a higher rate means RADR, probabilities mean expected NPV and standard deviation.
In MCQs, remember the logic: CE discounts at the risk-free rate, and a lower α means more risk is removed.
Show a small table for PV factors. Method marks are given even if one multiplication goes wrong.
For decision trees, draw the tree neatly and show EMV at every node. Roll back from the right.
State the recommendation in one line at the end, with the reason. In case-based questions, also mention the limitation of the method used, for example that the RADR applies the same rate to every year.

Practice questions from Investment Decisions, Project Planning and Control

Risk Analysis in Capital Budgeting in other exams

The same ground in other exams, if you are preparing for more than one or want another angle on it.

Risk Analysis in Capital Budgeting: frequently asked questions

What is the difference between certainty equivalent and risk-adjusted discount rate?

RADR keeps expected cash flows and raises the discount rate to allow for risk. CE reduces each cash flow to a sure amount using a coefficient and discounts at the risk-free rate. CE can show different risk for each year, while RADR uses one rate for all years.

How do I calculate the standard deviation of NPV?

For scenarios, find the NPV of each, then the expected NPV, then take the square root of the probability-weighted squared deviations. For yearly cash flows, use the discounted-variance formula if the flows are independent, or the sum of discounted standard deviations if they are perfectly correlated.

What is the difference between sensitivity analysis and scenario analysis?

Sensitivity analysis changes one variable at a time and shows its effect on NPV. Scenario analysis changes a combination of variables together to build cases such as pessimistic, most likely and optimistic.

When is a decision tree used in capital budgeting?

Use it when the project has stages and later decisions depend on earlier outcomes. You draw branches with probabilities and cash flows, then roll back from the end to choose the best first decision by expected NPV.

What does simulation add over sensitivity analysis?

Simulation varies many uncertain inputs at once using assumed probability distributions and repeats the calculation many times. The result is a distribution of NPV, from which you can read the chance of a loss. It needs more data and software, so exams usually ask for the concept and steps rather than a full numerical.