Capital Budgeting Under Risk and Constraints

Introduction

Capital budgeting is the process of evaluating long-term investments such as production facilities, technology systems, distribution centres, product launches and acquisitions. In earlier analysis, projects are often evaluated using expected cash flows and a discount rate. In practice, however, future cash flows are uncertain, capital is limited, projects may be interdependent, and managers may have flexibility to delay, expand, abandon or modify an investment.

A sound investment decision therefore requires more than calculating one net present value (NPV). It requires an explicit assessment of risk, resource constraints and managerial flexibility. This chapter develops practical methods for doing so: sensitivity analysis, scenario analysis, simulation, decision trees, real-options analysis, risk-adjusted discount rates, capital rationing, resolution of project-ranking conflicts and post-investment review.

Learning Objectives

By the end of this chapter, you should be able to:

1. Distinguish project risk from uncertainty and identify major sources of investment risk.
2. Apply sensitivity, scenario and simulation analysis to capital-budgeting decisions.
3. Construct and interpret decision trees using expected monetary value.
4. Explain real options and calculate simple option values.
5. Evaluate the strengths and limitations of risk-adjusted discount rates.
6. Allocate scarce capital when projects are indivisible or mutually exclusive.
7. Resolve conflicts between NPV, profitability index and other ranking methods.
8. Design an effective post-investment review and recognise governance and ethical issues.

1. Risk in Capital Investment

A project is risky when its actual outcome may differ from the outcome forecast at the time of approval. Risk may arise from:

• Market risk: changes in selling prices, demand, competition and exchange rates.
• Operating risk: changes in input prices, production volumes, labour productivity or operating costs.
• Technological risk: equipment failure, obsolescence, cybersecurity events or unsuccessful development.
• Financial risk: changes in interest rates, refinancing conditions and access to capital.
• Regulatory and political risk: taxation, environmental regulation, licensing and trade restrictions.
• Project and implementation risk: delays, cost overruns, supplier failure and inadequate project management.

Risk and uncertainty are sometimes distinguished conceptually. Risk refers to situations in which probabilities can be estimated, perhaps from historical data or expert analysis. Uncertainty refers to situations in which reliable probabilities are unavailable. In practice, the boundary is not precise; both require disciplined analysis and judgement.

The basic NPV model is:

Here, I₀ is the initial investment at time zero, CFₜ is the incremental after-tax cash flow in period t, r is the discount rate, n is the project life, and TVₙ is terminal value, if any. Cash flows must be incremental, after tax and measured consistently in nominal or real terms. A positive NPV means that, under the stated assumptions, the project is expected to increase shareholder wealth.

The risk problem arises because CFₜ, TVₙ and sometimes r are estimates rather than known quantities.

2. Sensitivity Analysis

Sensitivity analysis changes one important assumption at a time while holding other assumptions constant. It answers the question: “How sensitive is project value to a change in this variable?” Common variables include unit sales, selling price, variable cost, fixed cost, initial investment, project life and salvage value.

A simple sensitivity measure is the change in NPV resulting from a change in an input:

where X is the input variable. A higher absolute value indicates greater exposure to that variable, although variables measured in different units should not be compared without standardisation.

A useful percentage measure is the elasticity of NPV:

This ratio is dimensionless and indicates the percentage change in NPV associated with a 1% change in the input, within the tested range. Sensitivity analysis is most useful for identifying value drivers, break-even assumptions and information priorities.

Break-even analysis

Suppose annual unit sales are the uncertain variable. The operating break-even quantity, ignoring financing costs, is:

where Q_BE is break-even units, F is annual fixed operating cost, P is price per unit and V is variable cost per unit. This formula applies to a single-product setting with constant unit price and variable cost. It is not a substitute for NPV break-even analysis when cash-flow timing, tax, depreciation or terminal value is material.

An NPV break-even analysis instead solves for the value of an assumption that makes NPV equal to zero. For example, the minimum annual sales volume can be found by setting the complete after-tax NPV equation equal to zero and solving for quantity.

Worked example: one-way sensitivity

An investment of $1,000,000 generates an expected after-tax cash flow of $400,000 at the end of each of three years. The discount rate is 10% and there is no terminal value. The base NPV is:

The project is approximately value-neutral. If annual cash flow falls by 10% to $360,000, the present value of the annuity becomes approximately $895,263 and NPV becomes −$104,737. If annual cash flow rises by 10% to $440,000, NPV becomes approximately $94,211. The analysis shows that the investment is highly sensitive to annual operating cash flow.

A limitation is that changing one variable at a time may create unrealistic combinations. For instance, a lower selling price may normally be accompanied by higher volume, but one-way sensitivity holds volume constant.

3. Scenario Analysis

Scenario analysis changes several related assumptions simultaneously. Typical scenarios are pessimistic, base-case and optimistic. A scenario should be internally coherent: an economic downturn may reduce volume, selling price and input costs, while a strong economy may increase demand but also intensify labour and material costs.

If scenarios are assigned probabilities, expected NPV is:

where pⱼ is the probability of scenario j, NPVⱼ is its NPV and ∑pⱼ = 1. The variance of NPV is:

and the standard deviation is σ_NPV = √σ²_NPV. These measures are meaningful only if the scenarios are sufficiently comprehensive and probabilities are credible.

Scenario analysis is superior to one-way sensitivity when variables are correlated, but it can still understate risk if extreme outcomes are omitted. It should be used to support, not replace, managerial judgement.

4. Simulation Analysis

Simulation models uncertainty by assigning probability distributions to uncertain inputs and repeatedly calculating project cash flows and NPV. A Monte Carlo procedure typically follows these steps:

1. Identify uncertain inputs, such as demand, price, cost and exchange rate.
2. Specify probability distributions and relationships among inputs.
3. Draw a random value for each input.
4. Calculate cash flows and NPV for that trial.
5. Repeat the process many times.
6. Analyse the resulting distribution of NPV, including its mean, standard deviation and probability of a negative NPV.

Simulation produces a distribution rather than a single estimate. For example, management may learn that expected NPV is $2 million, but there is a 28% probability that NPV is negative. This is more informative than saying only that expected NPV is positive.

The quality of simulation depends on model structure, input distributions and correlations. Treating correlated variables as independent can produce misleadingly narrow results. Simulation also does not eliminate model risk: a sophisticated model can still be wrong if its assumptions are poor.

5. Decision-Tree Analysis

Decision trees represent sequential decisions and uncertain events. They are appropriate when management receives information over time and can respond. A tree contains decision nodes, chance nodes, probabilities and cash-flow outcomes.

The expected monetary value (EMV) at a chance node is:

where pᵢ is the probability of outcome i and Payoffᵢ is its payoff. Probabilities at a complete chance node must sum to one. Decision nodes are solved by selecting the alternative with the highest value, after discounting cash flows to a common date.

Illustrative decision tree

A company may spend $200,000 on a pilot project. If the pilot succeeds, with probability 0.6, the company can invest $800,000 in full production and receive a present-value benefit of $1,800,000. If it fails, the pilot has no further value. Assume the pilot and expansion amounts are expressed at the relevant decision dates and that discounting has already been incorporated for simplicity.

Value after a successful pilot is:

The expected value before the pilot is:

The pilot should be undertaken on these assumptions. Importantly, the company does not commit to full production after failure. This ability to abandon is the source of value that a conventional single-stage NPV may overlook.

Decision trees become difficult when there are many stages, continuous variables or feedback effects. They also require careful probability assessment and avoidance of double-counting risk in both probabilities and discount rates.

6. Real Options in Investment Decisions

A real option is the right, but not the obligation, to take a future business action. Unlike a financial option, it concerns a real asset or project. Common real options include:

• Option to delay: wait for more information before investing.
• Option to expand: increase capacity if demand is strong.
• Option to abandon: sell assets or stop a project if performance is poor.
• Option to contract: reduce scale if demand weakens.
• Option to switch: change inputs, outputs or production technology.
• Option to stage: invest progressively rather than all at once.

A conventional NPV often assumes a passive manager who follows a fixed plan. A real-options perspective recognises that flexibility has value. The expanded project value can be expressed conceptually as:

This is a conceptual relationship, not a universal valuation formula. Option value is generally higher when uncertainty is high, the time available to decide is long, the investment required to exercise the option is relatively low and the underlying project value is substantial.

Real-options analysis can use decision trees, binomial models or option-pricing methods. Financial option analogies must be applied carefully because real assets are often illiquid, non-traded and affected by managerial actions. Option value should not be used to justify every strategically attractive project; assumptions must remain transparent.

7. Risk-Adjusted Discount Rates

The risk-adjusted discount-rate (RADR) approach uses a higher discount rate for riskier projects. The general NPV formula remains:

The project discount rate should reflect the opportunity cost of capital for projects with comparable operating risk. A common framework is the capital asset pricing model (CAPM):

where r_e is the required return on equity, r_f is the risk-free rate, β_L is levered beta and r_m − r_f is the market risk premium. For a project, a comparable or unlevered beta may be used and then adjusted for the project’s target financing structure. A weighted average cost of capital may then be applied when project cash flows are appropriate for that rate:

Here, E and D are market values of equity and debt, r_d is the pre-tax cost of debt and T is the corporate tax rate. This formulation assumes stable financing proportions and that the tax shield is appropriately represented.

RADR is convenient, but it has important limitations. It assumes that risk can be summarised by one rate, which may be inappropriate when risk changes over time. It can also penalise long-term cash flows excessively and obscure the distinction between diversifiable and systematic risk. Adjusting cash flows for risk and then applying a high discount rate may double-count risk. Risk-adjusted cash-flow approaches, certainty equivalents and scenario or simulation methods can be useful alternatives.

8. Capital Rationing and Project Selection

Capital rationing occurs when the firm cannot undertake every positive-NPV project. The constraint may be a shortage of cash, borrowing capacity, skilled staff, production capacity, management attention or regulatory approval.

Profitability index

For an independent project with a single investment at time zero, the profitability index (PI) is:

Equivalently, when the initial investment is positive:

PI is useful under a single-period capital constraint because it measures value created per dollar invested. However, it should not automatically replace NPV. Projects may be indivisible, have different lives, require future investment or be mutually exclusive.

Worked capital-rationing example

Assume a firm has $1,000,000 available and the following independent projects are indivisible:

Selecting C and D uses $800,000 and creates $235,000 of NPV. Selecting A and B uses the full $1,000,000 and creates $280,000. Selecting B and C uses $900,000 and creates $260,000. Therefore A and B are the best combination among these alternatives, even though C has the highest individual PI. The correct procedure is to evaluate feasible combinations, especially when projects are indivisible.

If projects are divisible and the constraint is a single-period budget, ranking by PI may be effective. If constraints are multiple or extend across several periods, linear programming, integer programming or project portfolio optimisation may be required.

Ranking conflicts

NPV and internal rate of return (IRR) may rank mutually exclusive projects differently because of differences in scale, timing and reinvestment assumptions. NPV is generally preferred because it measures absolute wealth creation at the opportunity cost of capital. Incremental analysis can resolve conflicts: compare the incremental cash flows of the larger project with the smaller project and determine whether the incremental investment has a positive NPV.

Other conflicts arise from differing project lives. Equivalent annual value (EAV) converts an NPV into an annual equivalent for projects that can be repeated:

This assumes projects can be repeated under comparable conditions and that the discount rate remains appropriate. If repetition is unrealistic, replacement-chain analysis or a finite planning horizon may be preferable.

9. Post-Investment Review

A post-investment review, or post-completion audit, compares actual outcomes with the original investment case. It should examine:

• actual versus forecast investment expenditure;
• actual versus forecast sales volume, price and operating costs;
• timing of construction, launch and cash flows;
• changes in tax, regulation and competitive conditions;
• whether risk-management actions were effective; and
• whether the project remains economically justified.

The review should distinguish forecasting error from factors that were unforeseeable at approval. A useful variance framework is:

Variances should then be decomposed into price, volume, cost, timing and scope effects where possible. The purpose is organisational learning, not assigning blame. Reviews should be independent enough to be credible but conducted early enough to improve future decisions.

Post-investment review also supports capital reallocation. A project with a positive original NPV may now have a negative value if conditions change. Continuing solely because money has already been spent is the sunk-cost fallacy. Future decisions should be based on incremental future cash flows and available alternatives.

10. Integrated Illustrative Case

Background

Illustrative case: Northstar Foods is considering a $4 million automated packaging line. The line is expected to operate for five years, reduce labour costs and increase capacity. Demand for the new product is uncertain, and management has only $4.5 million available for all investments this year.

Problem

The base-case NPV is positive, but the project depends heavily on demand forecasts. A competing project, a warehouse-management system, has a smaller NPV but requires less capital. Northstar must assess risk and select an affordable portfolio.

Analysis

The finance team first performs sensitivity analysis. NPV is most sensitive to product volume and the implementation date. Scenario analysis produces three internally consistent cases:

Expected NPV is:

The probability of a negative NPV is 25%, which may be unacceptable under the company’s risk policy. Management then considers staging: a $0.5 million pilot can be completed before the full production commitment. If demand is weak, the project can be abandoned. This creates an abandonment option and reduces downside exposure.

The warehouse-management system requires $1.5 million and has an NPV of $0.55 million. If the packaging line is undertaken immediately, only $0.5 million remains, so the warehouse project cannot be funded. A staged packaging investment allows Northstar to preserve flexibility and potentially undertake both projects in a later period.

Approach and solution

The board approves the pilot subject to:

1. a defined demand threshold for full investment;
2. an independent review of the pilot results;
3. a maximum approved cost and procurement controls;
4. a simulation report showing the probability of negative NPV; and
5. a post-investment review twelve months after launch.

The decision does not rely on expected NPV alone. It combines scenario analysis, a real option, capital constraints and governance safeguards.

Learning points

• Expected NPV can conceal substantial downside risk.
• Staging may be valuable when information arrives over time.
• Capital rationing requires portfolio analysis rather than isolated project acceptance.
• Decision rights and review thresholds should be specified before implementation.
• A post-investment audit improves future forecasts and accountability.

11. Governance, Ethics and Technology

Investment proposals create incentives to overstate benefits, understate costs or present optimistic scenarios to obtain approval. Ethical capital budgeting requires honest assumptions, clear disclosure of uncertainty, consistent treatment of alternatives and protection against conflicts of interest. Approval committees should challenge strategic enthusiasm and require evidence for probability estimates.

Useful controls include segregation of duties, documented approval limits, sensitivity thresholds, independent technical review and audit trails for model changes. Environmental and social effects should be considered where they create economic, legal or reputational consequences. A project should not be labelled financially attractive by excluding material external costs that the organisation is likely to bear.

Spreadsheet models and artificial-intelligence tools can accelerate simulation, scenario generation and anomaly detection. They do not replace professional judgement. Users should validate data, inspect formulas, test extreme assumptions, document model versions and avoid treating machine-generated forecasts as objective facts. Explainability and human approval are especially important for high-value or irreversible investments.

12. Best-Practice Decision Process

A disciplined process can be summarised as follows:

1. Define the project, alternatives and decision horizon.
2. Forecast incremental after-tax cash flows, including working capital and terminal effects.
3. Identify key uncertainties and dependencies.
4. Apply sensitivity and coherent scenario analysis.
5. Use simulation when distributions and correlations materially affect the decision.
6. Use decision trees or real-options analysis when decisions occur sequentially.
7. Select a discount rate consistent with project risk and financing assumptions.
8. Apply capital-rationing and portfolio analysis under resource constraints.
9. Document assumptions, approvals, risks and decision thresholds.
10. Conduct a post-investment review and update organisational forecasting methods.

Chapter Summary

Capital budgeting under risk recognises that project cash flows, timing and strategic conditions are uncertain. Sensitivity analysis identifies individual value drivers; scenario analysis examines coherent combinations of assumptions; and simulation estimates a distribution of possible NPVs. Decision trees are useful when uncertainty is resolved over time and management can respond. Real-options analysis values flexibility such as delaying, expanding, staging or abandoning an investment.

Risk-adjusted discount rates provide a convenient method for incorporating risk, but they must reflect comparable project risk and should not double-count uncertainty. Under capital rationing, profitability indexes can assist preliminary ranking, but feasible project combinations should be evaluated when investments are indivisible. NPV remains the primary wealth-creation criterion, particularly when projects differ in scale or timing. Post-investment review promotes accountability, improves forecasts and prevents sunk-cost reasoning. Effective capital budgeting combines quantitative analysis, governance, ethical judgement and managerial flexibility.

Key Terms

Risk: Possibility that actual project outcomes differ from forecasts.

Uncertainty: Lack of reliable knowledge about future outcomes or probabilities.

Sensitivity analysis: Testing the effect of changing one assumption while holding others constant.

Scenario analysis: Evaluating project value under internally consistent sets of assumptions.

Simulation: Repeated calculation of outcomes using randomly drawn values from probability distributions.

Monte Carlo simulation: Simulation method that uses repeated random sampling.

Decision tree: Diagram of sequential decisions, chance events and outcomes.

Expected monetary value: Probability-weighted average of possible payoffs.

Real option: Right, but not obligation, to take a future business action.

Option to abandon: Flexibility to terminate a project or sell its assets.

Risk-adjusted discount rate: Discount rate increased or adjusted to reflect project risk.

Capital rationing: Situation in which available investment resources are limited.

Profitability index: Present value of future inflows divided by initial investment.

Incremental cash flow: Cash flow caused by accepting a project compared with not accepting it.

Equivalent annual value: Annualised value of an investment’s NPV.

Post-investment review: Comparison of actual project results with the approved forecast.

Model risk: Risk that a model is incorrectly structured or based on unsuitable assumptions.

Sunk cost: Past expenditure that cannot be changed by a current decision.

Review Questions

5 Short Answer Questions

1. What is the main purpose of sensitivity analysis?
2. State two situations in which decision-tree analysis is useful.
3. Define a real option and give one example.
4. Why can profitability-index rankings be misleading under capital rationing?
5. What is the purpose of a post-investment review?

5 Long Answer Questions

1. Compare sensitivity analysis, scenario analysis and simulation, including the strengths and limitations of each.
2. Explain how a risk-adjusted discount rate can be estimated and discuss the dangers of using one rate for all projects.
3. Describe how decision trees and real-options analysis improve upon a passive, single-stage NPV analysis.
4. Explain why NPV is generally preferred to IRR when mutually exclusive projects have different scales or timing patterns.
5. Discuss the governance and ethical controls that should surround major capital-investment decisions.

3 Analytical / Application Questions

1. A project requires $2 million and has possible NPVs of −$0.5 million, $0.8 million and $1.6 million with probabilities of 0.20, 0.50 and 0.30. Calculate expected NPV and the probability of a negative NPV. Interpret the result.
2. A firm has $1 million available and three indivisible projects: A requires $600,000 and has NPV $180,000; B requires $400,000 and has NPV $100,000; C requires $500,000 and has NPV $160,000. Identify the best feasible combination and explain why individual profitability-index ranking is insufficient.
3. Design a post-investment review for a technology project whose implementation was delayed and whose operating savings are below forecast. Specify the variances to investigate and explain how the findings should affect future decisions.

Multiple Choice Questions with Answers

Question 1

Which method changes one input assumption at a time while holding other assumptions constant?

A. Sensitivity analysis

B. Scenario analysis

C. Monte Carlo simulation

D. Decision-tree analysis

Correct Answer: A. Sensitivity analysis

Question 2

A project has NPVs of −$100,000 and $300,000 under two equally likely scenarios. What is its expected NPV?

A. −$200,000

B. $100,000

C. $200,000

D. $400,000

Correct Answer: B. $100,000

Question 3

Which feature most directly creates value in a real option to abandon?

A. The ability to avoid selected future losses

B. The certainty of higher initial investment

C. The elimination of all project uncertainty

D. The requirement to continue after poor results

Correct Answer: A. The ability to avoid selected future losses

Question 4

At a chance node, expected monetary value is calculated by:

A. Adding all payoffs without probabilities

B. Multiplying each payoff by its probability and summing

C. Selecting the smallest payoff

D. Discounting only the initial investment

Correct Answer: B. Multiplying each payoff by its probability and summing

Question 5

If a firm has a single-period budget and projects are divisible, which measure can help rank projects by value per dollar invested?

A. Accounting rate of return

B. Profitability index

C. Payback period only

D. Debt-to-equity ratio

Correct Answer: B. Profitability index

Question 6

Why can NPV and IRR rank mutually exclusive projects differently?

A. They always use different tax rates

B. They respond differently to project scale and cash-flow timing

C. NPV ignores cash flows after year one

D. IRR cannot be calculated for any project

Correct Answer: B. They respond differently to project scale and cash-flow timing

Question 7

If a project requires $500,000 and has an NPV of $100,000, its profitability index is:

A. 0.20

B. 0.80

C. 1.20

D. 5.00

Correct Answer: C. 1.20

Question 8

Which statement about a risk-adjusted discount rate is most accurate?

A. The same rate is always appropriate for every project

B. It should reflect the risk of comparable project cash flows

C. It eliminates the need to forecast cash flows

D. It guarantees that NPV will be positive

Correct Answer: B. It should reflect the risk of comparable project cash flows

Question 9

A post-investment review should primarily:

A. Ignore forecast errors caused by timing

B. Compare actual results with approved forecasts and identify causes

C. Approve every project that has already incurred spending

D. Replace all future cash-flow forecasts with historical data

Correct Answer: B. Compare actual results with approved forecasts and identify causes

Question 10

A project costs $1,000,000 and produces $400,000 at the end of each year for three years. At a 10% discount rate, its NPV is approximately:

A. −$5,263

B. $5,263

C. $194,737

D. $400,000

Correct Answer: A. −$5,263