A forecast becomes dangerous when its precision hides how uncertain the underlying future really is.
A capital proposal can contain hundreds of spreadsheet lines and still present only one future. Revenue grows at a specified rate. Capital expenditure lands on schedule. Input costs follow the assumed curve. Benefits arrive when planned. The model may be internally consistent and mathematically correct, yet the decision can remain poorly informed because the leadership team sees the base case more clearly than the uncertainty around it.
The investment-risk material supplied for MPM416 starts from a simple proposition: investment decisions concern future events, and future events are inherently uncertain. Hargitay and Yu define risk in terms of the extent to which actual outcomes may diverge from expected outcomes. The strategic implication is significant. Risk is not adequately represented by a single premium, traffic-light rating or contingency percentage. It is better understood as a range and pattern of possible outcomes.
The Strategic Context
Modern investment appraisal handles the time value of money well. Discounted cash flow techniques can translate future cash flows into present values. The harder problem is that future cash flows are not known with certainty.
Revenue may depend on market adoption. Construction cost may depend on commodity prices, design maturity and productivity. Technology performance may depend on integration. Benefits may depend on behavioural adoption. Regulatory conditions may change. Several variables may move together rather than independently.
That uncertainty matters because two projects with the same expected NPV can have very different downside exposures. One may have a narrow range around the expected value. Another may combine a large upside with a material probability of capital loss. A single expected number conceals that difference.
McKinsey's supplied capital-project article gives this idea a practical executive form. It advocates describing a project's value through a probability distribution, key risk metrics and explicit sources of risk. The point is not mathematical sophistication for its own sake. It is to improve the discussion about possible outcomes and trade-offs.
What Leaders Commonly Misread
The first misreading is base-case anchoring. Once a business case contains a central forecast, subsequent debate tends to orbit around it. Even when downside and upside cases exist, the base case often becomes psychologically privileged.
The second is false objectivity. A model may calculate to two decimal places while its most important assumptions are based on judgement. Precision in calculation does not increase precision in the assumptions.
The third is risk-premium compression. Organisations sometimes respond to uncertainty by increasing the discount rate. That may be appropriate in some contexts, but it can also hide the sources of risk. A larger discount rate does not tell leaders whether the threat is capital overrun, schedule delay, demand volatility, regulatory exposure or technology failure. Different risks require different responses.
The fourth is to assume that historical data can forecast structural change. The supplied Hargitay/Yu material recognises the usefulness of past distributions in the short run while also warning that predictive precision deteriorates rapidly. Historical evidence is an input to judgement, not a guarantee that the future will resemble the past.
Reframing the Issue
The core decision question should move from 'What is the return?' to 'What outcomes are plausible, how likely are they, what drives the spread, and which outcomes can we influence?'
This matters because risk analysis should change action. If the major downside driver is design immaturity, leadership can fund additional engineering before sanction. If the problem is demand uncertainty, it can stage capacity. If the threat is a correlated market exposure across the portfolio, it can reduce concentration or hedge. If downside cannot be economically mitigated, the organisation can decline the investment.
Risk analysis that does not change design, timing, governance or portfolio composition is often just documentation.
Strategic Analysis: Four Levels of Uncertainty Treatment
Sensitivity analysis: identify what matters
Sensitivity analysis changes one factor and observes the effect on the projected outcome. The supplied study notes emphasise that sensitivity analysis is particularly useful for identifying variables that could materially influence success or failure. Its strength is interpretability.
The weakness is that variables rarely move one at a time in real life. Sensitivity analysis is therefore best treated as a diagnostic tool: it tells leaders where deeper analysis is justified.
Scenario analysis: test coherent futures
Scenarios change several assumptions together to describe internally coherent futures. A downturn can combine lower demand, delayed revenue, tighter financing and pressure on working capital. A regulatory scenario can combine approval delay, redesign cost and market-entry timing.
The value of scenarios is not prediction. It is preparedness. They reveal which strategies remain viable across different futures and which depend on a narrow set of conditions.
Probability analysis: show the shape of outcomes
Where credible probability assumptions can be developed, a distribution of NPV, cash flow or return can show expected value, downside percentiles, break-even probability and the likelihood of achieving the baseline.
This can change executive dialogue. A project may have an attractive expected value but only a modest chance of achieving the sponsor's base case. Conversely, a project with a lower headline return may have a much more resilient downside profile.
Simulation: examine interacting uncertainty
The Hargitay/Yu material describes probabilistic simulation as a powerful way to examine complex investment problems when multiple variable factors interact. The quality of the result, however, depends on the quality of the model and the relationships assumed between inputs.
Simulation should therefore never be presented as a machine for producing truth. It is a way to explore the consequences of a model of reality. If the model is poor, the simulation can produce impressive-looking nonsense at scale.
Decision Framework
Executives do not need to choose the most sophisticated technique. They need the minimum sufficient analysis for the decision's scale, irreversibility and uncertainty.
| Decision condition | Appropriate treatment |
|---|---|
| Few dominant uncertain variables | Sensitivity analysis |
| Distinct plausible future states | Scenario analysis |
| Important distribution of financial outcomes | Probability-based analysis |
| Multiple interacting uncertain variables | Simulation |
| High portfolio concentration | Portfolio/correlation analysis |
Five governance questions should accompany the numbers.
First, which assumptions have the greatest effect on value? Second, which assumptions are evidence-based and which are judgement? Third, which risks are correlated with risks already in the portfolio? Fourth, what mitigation changes the distribution materially rather than cosmetically? Fifth, what trigger would cause the investment decision to be revisited?
A useful board-level output is not a dense model. It is a one-page uncertainty profile showing the expected outcome, credible downside, major value drivers, assumptions with weak evidence, mitigation options and decision triggers.
Related article: From Project Economics to Portfolio Choice: Why NPV Is Not Enough
From Strategy to Execution
Immediately, stop allowing large investment proposals to present only one forecast. At minimum, require sensitivity to the handful of variables that matter most.
In the medium term, standardise the treatment of uncertainty across comparable projects. Common conventions for inflation, schedule risk, contingency, demand scenarios and cost ranges make portfolio comparison more credible.
For major irreversible investments, establish independent challenge. The people who developed a proposal are usually deeply knowledgeable, but they may also be anchored to the proposal's success. Independent review should test assumptions, not merely recalculate formulas.
Longer term, connect risk analysis to benefits realisation and organisational learning. When actual outcomes become available, compare them with the ranges assumed at approval. Were cost distributions too narrow? Was adoption systematically overestimated? Did certain risks recur? This closes the loop between forecasting and future investment quality.
Signals to Monitor
Warning signs include narrow forecast ranges despite weak evidence; repeated use of the same contingency percentage regardless of project type; risks expressed as labels without financial or operational consequences; models where downside variables are assumed independent despite common drivers; and post-investment reviews that compare actuals only with the base case rather than the original range of expected outcomes.
Another signal is behavioural: when executives ask project teams to 'make the numbers work', the organisation is no longer analysing uncertainty. It is negotiating the forecast.
Questions for the Leadership Team
- Which three assumptions account for most of the variation in the value of this investment?
- What is the credible downside, not merely the expected outcome?
- Which risks can management actually influence before or after commitment?
- Where are we using historical stability to justify a future that may be structurally different?
- Which project risks are correlated with exposures already present elsewhere in the portfolio?
- What evidence would cause us to change the probability assumptions or stop the project?
Closing Perspective
The goal of risk analysis is not to eliminate uncertainty. That is impossible. The goal is to prevent uncertainty from being hidden inside an apparently precise business case.
Better decisions emerge when leaders can see the range of plausible outcomes, understand what creates that range and decide which exposures are worth carrying. A single number can support a decision. A distribution helps leaders understand the decision they are actually making.
Source Foundations
- Hargitay, S. E. & Yu, S.-M., “The analysis of investment risk”, supplied university extracts from Property Investment Decisions.
- Pergler, M. & Rasmussen, A., “Making better decisions about the risks of capital projects”, McKinsey & Company, May 2014.