Why Risk Belongs in a Strategic Choice
A growth option can look attractive because its best-case outcome is large, because its average outcome is favorable, or because it advances a strategic objective. Those are different reasons, and none alone establishes that the option is appropriate. A sound decision considers what the organization is trying to achieve, the range of possible outcomes, and whether the adverse outcomes are acceptable.
An enterprise risk management perspective brings those considerations into the choice itself. It does not mean rejecting every uncertain opportunity or reducing the decision to one numerical score. It means comparing alternatives on a consistent basis, identifying how uncertainty could affect the objectives, and making explicit which trade-offs the organization is willing to accept.
Build a Comparable Decision
Before calculating or comparing outcomes, define the decision. State the alternatives, the objective or objectives they are intended to support, and any risk constraints that cannot be exceeded. Then use a common horizon and valuation basis. Comparing one option’s annual revenue with another option’s five-year net contribution, for example, would not provide a meaningful comparison.
For a simple scenario analysis, let \(X\) denote an alternative’s net contribution over the chosen horizon. The possible values of \(X\) are its scenario outcomes, and each outcome has an assigned probability. The scenarios should represent plausible conditions that matter to the decision; their probabilities must be nonnegative and sum to 1. The figures and probabilities below are hypothetical. In each example, dollar amounts are stated in thousands.
Worked Comparisons
Worked Example: Comparing Two Growth Paths
A hypothetical insurer is choosing between expanding its digital distribution and purchasing a small regional agency. Management estimates each option’s five-year net present contribution, in $000, using a common valuation basis and an 8% effective annual discount rate. The estimates include initial investment and operating costs. For both options, the scenario probabilities are 0.25 for strong growth, 0.50 for the central case, and 0.25 for a severe slowdown.
| Option | Strong growth | Central case | Severe slowdown |
|---|---|---|---|
| Digital distribution | 30,000 | 12,000 | -8,000 |
| Agency purchase | 20,000 | 11,000 | 2,000 |
Method. Calculate each option’s expected contribution and variance using the same scenarios and probabilities, then compare the results with the organization’s stated risk constraint. Suppose the insurer requires expected contribution of at least $10,000 and a probability of negative contribution no greater than 10%.
For digital distribution, \(E[X]=0.25(30{,}000)+0.50(12{,}000)+0.25(-8{,}000)=7{,}500+6{,}000-2{,}000=11{,}500\). Its variance is \(0.25(30{,}000-11{,}500)^2+0.50(12{,}000-11{,}500)^2+0.25(-8{,}000-11{,}500)^2\) \(=0.25(18{,}500)^2+0.50(500)^2+0.25(-19{,}500)^2\) \(=85{,}562{,}500+125{,}000+95{,}062{,}500=180{,}750{,}000\). The probability of a negative contribution is 0.25, since only the severe-slowdown outcome is negative.
For the agency purchase, \(E[X]=0.25(20{,}000)+0.50(11{,}000)+0.25(2{,}000)=5{,}000+5{,}500+500=11{,}000\). Its variance is \(0.25(20{,}000-11{,}000)^2+0.50(11{,}000-11{,}000)^2+0.25(2{,}000-11{,}000)^2\) \(=0.25(9{,}000)^2+0+0.25(-9{,}000)^2=20{,}250{,}000+20{,}250{,}000=40{,}500{,}000\). None of its scenario outcomes is negative, so its probability of negative contribution is 0.
Both options exceed the expected-contribution minimum. The agency purchase has a lower expected contribution by $500,000, but it meets the downside constraint; digital distribution does not. Under the stated constraint, the agency purchase is the admissible choice, subject to other strategic considerations. This comparison does not prove that it is universally better.
Check and common mistake avoided. The three probabilities sum to 1 for each option, and the expected values are weighted sums of the stated outcomes. A common mistake is to select digital distribution because its expected contribution is higher, without testing the risk constraint. That ignores a decision condition that management has explicitly set.
Worked Example: Separating a Strategic Objective From a Financial Constraint
A hypothetical insurer considers two ways to enter an adjacent market. Both advance the objective of establishing a presence there, but the first option would reach more potential customers if successful. The five-year net contributions, in $000, are estimated on the same basis, with scenario probabilities of 0.20 for strong demand, 0.60 for moderate demand, and 0.20 for weak demand.
Option A has outcomes of 12,000, 8,000, and -4,000. Its expected contribution is \(0.20(12{,}000)+0.60(8{,}000)+0.20(-4{,}000)=2{,}400+4{,}800-800=6{,}400\). The probability of a negative contribution is 0.20.
Option B has outcomes of 9,000, 7,000, and 3,000. Its expected contribution is \(0.20(9{,}000)+0.60(7{,}000)+0.20(3{,}000)=1{,}800+4{,}200+600=6{,}600\). All its scenario outcomes are positive, so its probability of a negative contribution is 0.
Suppose the organization has set a minimum expected contribution of $6,500 and a maximum probability of negative contribution of 10%. Option A advances the market-reach objective more strongly, but it meets neither financial condition. Option B meets both conditions and still supports entry into the adjacent market. On the information provided, B fits the stated constraints better; management would need to decide whether A’s additional reach justifies reconsidering those constraints.
Check and common mistake avoided. The expected contributions are $6,400 and $6,600, respectively, while the downside probabilities are 20% and 0%. Do not treat the strategic objective as a substitute for financial and risk conditions, or assume that satisfying those conditions automatically resolves every strategic trade-off.
Why the Weighted-Average Formula Works
In a discrete scenario model, \(X\) takes exactly one of the listed outcomes. The expected value is the long-run average contribution per repetition of the same uncertain decision, if the scenario probabilities correctly describe the uncertainty and the setting is repeated under comparable conditions. Each outcome contributes to that average in proportion to how often its scenario is expected to occur.
Variance follows the same logic, but applies it to squared deviations from the expected value. If the outcome is \(x_j\), its deviation is \(x_j-E[X]\); squaring makes negative and positive deviations contribute positively. Weighting those squared deviations by their scenario probabilities gives the average squared spread:
This derivation also shows a limitation: variance measures dispersion, not whether the dispersion is harmful to a particular organization. Two options with equal variance can have different probabilities of falling below a critical threshold, and a financially adverse outcome may matter differently depending on the organization’s objectives and capacity to absorb it. Use measures that directly address the decision’s stated constraints as well as summary measures such as expected contribution and variance.
Worked Example: Testing a Key Probability Assumption
A hypothetical insurer is considering a new product. If regulatory approval is obtained, the product’s five-year net present contribution is estimated at $24,000; if approval is denied, the contribution is -$6,000. These amounts are in $000 and use the same 8% effective annual discount rate and five-year horizon. A separate, lower-growth alternative has a certain contribution of $8,000. Let \(p\) be the probability of approval.
Method. Compare the expected contribution of the approval-dependent product with the certain alternative, and solve for the probability at which the two expected contributions are equal: \(E[X]=24{,}000p+(-6{,}000)(1-p)=30{,}000p-6{,}000\).
Set this equal to $8,000: \(30{,}000p-6{,}000=8{,}000\), so \(30{,}000p=14{,}000\), and \(p=14{,}000/30{,}000=0.4667\) to four decimal places.
At an estimated approval probability of 0.60, the expected contribution is \(30{,}000(0.60)-6{,}000=18{,}000-6{,}000=12{,}000\). The variance at this probability is \(0.60(24{,}000-12{,}000)^2+0.40(-6{,}000-12{,}000)^2\) \(=0.60(12{,}000)^2+0.40(-18{,}000)^2=86{,}400{,}000+129{,}600{,}000=216{,}000{,}000\). The probability of a negative contribution is 0.40. Thus the product has a higher expected contribution than the certain alternative at \(p=0.60\), but still exposes the organization to a material chance of a negative result.
Check and common mistake avoided. Substituting \(p=0.4667\) into the expected contribution gives approximately \(30{,}000(0.4667)-6{,}000=8{,}001\); the small difference from $8,000 is due to rounding. The exact break-even probability is \(14/30\). A common error is to multiply the failure outcome by \(p\) rather than by \(1-p\), reversing the approval and denial probabilities.
Use the Analysis to Make the Trade-Off Explicit
The calculations support a strategic decision; they do not make it automatically. A disciplined comparison can be organized as follows:
Specify the strategic outcome the alternatives are intended to support, along with measurable conditions that must be met.
Use a common time horizon, valuation basis, and definition of net contribution. State material assumptions, including the probabilities assigned to scenarios.
Calculate expected contributions and appropriate measures of downside or dispersion. Identify which scenario or assumption is responsible for a material adverse result.
Check each option against the organization’s objectives and risk constraints. Distinguish a binding constraint from a preference that may be weighed against other benefits.
Explain why the selected alternative fits the objectives, what trade-off was accepted, and which assumptions are important to the decision.
Scenario probabilities are estimates, not facts. If the choice changes materially when a probability shifts slightly, the decision is sensitive to that assumption. Decision-makers can then ask whether the estimate is sufficiently supported, whether a more conservative case changes the choice, or whether the organization is willing to accept the resulting exposure. Sensitivity analysis does not remove uncertainty; it makes the dependence of the decision on uncertainty visible.
A recurring exam trap is choosing the option with the largest best-case result or expected value without checking whether its downside conflicts with the objective or risk constraint. Another is comparing alternatives built on different assumptions or time horizons. A robust conclusion states both the result of the comparison and the condition that supports it: for example, “Option B meets the stated downside limit, whereas Option A does not,” rather than “Option B is safer” without defining the comparison.
Check Your Understanding
Use the stated scenario probabilities and answer each question. Monetary amounts are in $000.
- A proposed expansion has contributions of 14,000, 5,000, and -7,000 with probabilities 0.20, 0.50, and 0.30. Calculate its expected contribution and probability of a negative contribution. (Answer: Expected contribution = $3,200; probability of a negative contribution = 0.3000.)
- Option M has outcomes 10,000, 6,000, and 2,000; option N has outcomes 13,000, 5,000, and -3,000. For each, the probabilities are 0.25, 0.50, and 0.25. If the objective is to maximize expected contribution subject to a maximum 10% probability of negative contribution, which option qualifies? (Answer: M; its expected contribution is $6,000 and its probability of a negative contribution is 0, while N has expected contribution $5,000 and negative-outcome probability 0.2500.)
- A project yields 19,000 if successful and -5,000 otherwise. A certain alternative yields 6,000. What success probability makes their expected contributions equal? (Answer: \(p=11/24=0.4583\) to four decimal places.)
- A growth option has outcomes 8,000, 4,000, and -4,000 with probabilities 0.25, 0.50, and 0.25. Calculate its expected contribution and variance. (Answer: Expected contribution = $3,000; variance = 19,000,000 in squared $000 units.)
- Two alternatives have outcomes A: 15,000, 6,000, and -9,000; B: 9,000, 5,000, and 1,000, each with probabilities 0.20, 0.50, and 0.30. If the organization requires expected contribution of at least $4,000 and a probability of negative contribution no greater than 10%, which alternative meets both conditions? (Answer: B; its expected contribution is $4,600 and its probability of negative contribution is 0, while A has expected contribution $3,300 and negative-outcome probability 0.3000.)