Staged Investment Under Uncertainty

Why Staging Changes an Investment Decision

A company may be able to fund a small initial stage, receive information about demand or technical performance, and then decide whether to make a much larger investment. The initial spending buys more than a project: it also gives management a choice. If conditions are favorable, management can continue; if they are unfavorable, it can stop and avoid the later cost.

That choice can make a staged project more valuable than a plan that commits to every stage in advance. It does not, however, make every staged investment worthwhile. The initial stage still has a cost, and the value of flexibility must be measured against that cost. The key calculation is made after resolving uncertainty, state by state, rather than by averaging outcomes first.

Definition: A staged investment commits capital in steps. After an initial investment, management receives information and may choose whether to pay a later continuation cost. A project is flexible when management can decline that later investment; it has mandatory continuation when the later investment must be made regardless of the information received.

Use the following one-period setup. At time 0, the company pays an initial cost $C_0$. At time 1, it observes state $s$. In that state, the continuation project has value $V_s$, measured at time 1 before the continuation cost $K$ is paid. Continuing therefore produces net value $V_s-K$ at time 1. If management can stop, it receives zero from the continuation stage. The initial cost is already paid and cannot be recovered at the time-1 decision.

Valuation convention: Let $q_s$ be the valuation weight for state $s$, with $\sum_s q_s=1$. Then the flexible NPV is $ -C_0+\frac{1}{1+i}\sum_s q_s\max(V_s-K,0) $. The mandatory-continuation NPV is $ -C_0+\frac{1}{1+i}\sum_s q_s(V_s-K) $. Here $i$ is the effective annual rate appropriate to the valuation convention. In the worked examples, the weights are risk-neutral valuation probabilities and $i$ is the effective annual risk-free rate. With real-world probabilities, a risk adjustment is required; it is not generally valid to combine unadjusted probabilities with a risk-free rate.

These formulas assume a single time-1 decision, no additional cash flows between times 0 and 1, and no salvage value or other payoff from stopping. They also assume that $V_s$ and $K$ are measured at the same date and in the same units. If the continuation project has further stages, taxes, financing side effects, or a different risk profile, those cash flows and risks must be incorporated consistently rather than hidden in this simple one-period setup.

How the Flexibility Value Arises

The operation $\max(V_s-K,0)$ represents the decision available after state $s$ is known. Management continues when $V_s>K$ and stops when $V_s<K$. If $V_s=K$, either choice has zero incremental value under this model.

The reason the maximum must be taken before averaging is that the decision can differ across states. Starting with the mandatory-continuation value and comparing it with the flexible value gives:

\operatorname{NPV}_{\text{flex}}-\operatorname{NPV}_{\text{mandatory}} = \frac{1}{1+i}\sum_s q_s\left[\max(V_s-K,0)-(V_s-K)\right] = \frac{1}{1+i}\sum_s q_s\max(K-V_s,0).

For each state, if $V_s\ge K$, the bracketed difference is zero: management continues, so flexibility changes nothing in that state. If $V_s<K$, mandatory continuation loses $K-V_s$, while management with the option to stop avoids that loss. Thus the flexibility premium is nonnegative under these assumptions and equals the discounted valuation-weighted loss avoided by not continuing in unfavorable states.

Set the timeline.
Identify when the initial cost is paid, when information arrives, and when the continuation cost and value are measured.

Calculate each state’s continuation net value.
For each state, subtract the time-1 continuation cost $K$ from the time-1 value $V_s$.

Apply the decision rule.
For flexible investment, keep positive net values and replace negative values with zero. For mandatory continuation, keep all signed net values.

Average, discount, and subtract the initial cost.
Use the stated valuation weights, discount the time-1 amount to time 0, then subtract $C_0$.

Worked Valuations

Worked Example: Flexible Versus Mandatory Continuation

A hypothetical manufacturer pays a pilot cost of $1.00 million now. In one year, the valuation weights for three demand states are 0.20, 0.50, and 0.30. The corresponding time-1 continuation values are $14.00 million, $9.00 million, and $5.00 million. Continuing costs $9.00 million at time 1. Assume these are risk-neutral valuation weights and the effective annual risk-free rate is $i=8\%$. Find flexible NPV, mandatory-continuation NPV, and the flexibility premium.

Method. Calculate the net continuation value in each state, apply the stop-or-continue rule state by state, and discount the resulting weighted value. All monetary amounts below are in millions of dollars.

The state-by-state net values are $14.00-9.00=5.00$, $9.00-9.00=0.00$, and $5.00-9.00=-4.00$. With flexibility, management continues in the first state, is indifferent in the second, and stops in the third. The weighted time-1 flexible value is $0.20(5.00)+0.50(0.00)+0.30(0.00)=1.00$.

Discounting and subtracting the pilot cost gives $\operatorname{NPV}_{\text{flex}}=1.00/1.08-1.00=0.925926-1.00=-0.074074$, or −$0.0741 million to four decimal places.

With mandatory continuation, retain the negative value in the third state. The weighted time-1 value is $0.20(5.00)+0.50(0.00)+0.30(-4.00)=1.00-1.20=-0.20$. Therefore, $\operatorname{NPV}_{\text{mandatory}}=-0.20/1.08-1.00=-0.185185-1.00=-1.185185$, or −$1.1852 million to four decimal places. The flexibility premium is $-0.074074-(-1.185185)=1.111111$, or $1.1111 million to four decimal places. This equals the discounted avoided loss, $0.30(4.00)/1.08=1.111111$.

Check and common mistake. The flexible NPV is negative, so flexibility does not make this pilot acceptable at its stated cost. A tempting but incorrect calculation is to average the signed net values first, obtain $-0.20$, and then replace that average with zero because management can stop. That would give NPV of $-$1.00$ million. It is wrong because management decides after seeing the state: the correct flexible time-1 average is $1.00$, not zero. The state-by-state calculation also confirms that the flexibility premium is nonnegative.

Worked Example: Finding the Maximum Pilot Cost

A hypothetical energy-storage company considers a pilot costing $C_0$ now. In one year, the high state has valuation weight 0.40 and a continuation value of $11.00 million; the low state has weight 0.60 and a continuation value of $6.00 million. Continuing costs $8.00 million at time 1. Assume the weights are risk-neutral and $i=5\%$. Find the maximum pilot cost that gives nonnegative flexible NPV. Then find mandatory-continuation NPV at that pilot cost.

Method. First value the right to continue selectively. The time-1 net values are $11.00-8.00=3.00$ million in the high state and $6.00-8.00=-2.00$ million in the low state. Under flexibility, the company continues only in the high state. The expected time-1 flexible value is $0.40(3.00)+0.60(0)=1.20$ million.

The present value of that amount is $1.20/1.05=1.142857$ million. Flexible NPV is nonnegative precisely when $C_0\le1.142857$ million. The maximum pilot cost is therefore $1.1429 million, rounded to four decimal places in millions.

For mandatory continuation, the weighted time-1 net value is $0.40(3.00)+0.60(-2.00)=1.20-1.20=0$. At the maximum pilot cost, mandatory-continuation NPV is $0/1.05-1.142857=-1.142857$ million, or −$1.1429 million to four decimal places.

Check and common mistake. The mandatory time-1 value is zero because the favorable and unfavorable state contributions exactly offset. Do not count the high-state value of $3.00 million again as a benefit under mandatory continuation while also ignoring the low-state loss. The same cost $K$ is applied in each state, and the weighted net values must be computed before discounting.

Worked Example: Comparing Immediate and Staged Investment

A hypothetical medical-device firm can either invest $4.00 million immediately or pay $0.50 million for a pilot now and decide in one year whether to invest $4.00 million. At time 1, the continuation asset value is $7.00 million in the high state and $2.00 million in the low state; each state has valuation weight 0.50. Assume the weights are risk-neutral and $i=10\%$. The immediate investment costs $4.00 million at time 0 and yields the relevant asset value at time 1. Find the NPV of immediate investment, the flexible staged NPV, and the staged NPV if continuation is mandatory.

Method. For immediate investment, discount the weighted asset value and subtract the time-0 investment. The expected time-1 asset value is $0.50(7.00)+0.50(2.00)=4.50$ million. Thus immediate NPV is $4.50/1.10-4.00=4.090909-4.00=$ $0.0909 million, rounded to four decimal places in millions.

For the staged plan, the net continuation values at time 1 are $7.00-4.00=3.00$ million and $2.00-4.00=-2.00$ million. With flexibility, continue only in the high state, giving weighted time-1 value $0.50(3.00)+0.50(0)=1.50$ million. Staged flexible NPV is $1.50/1.10-0.50=1.363636-0.50=$ $0.8636 million.

If continuation is mandatory, the weighted time-1 net value is $0.50(3.00)+0.50(-2.00)=0.50$ million. Mandatory-continuation NPV is $0.50/1.10-0.50=0.454545-0.50=$ −$0.0455 million.

Check and common mistake. The staged flexible NPV exceeds the immediate NPV by $0.863636-0.090909=0.772727$ million. This comparison reflects both the different timing and size of the initial spending and the later choice; it is not just the flexibility premium relative to mandatory continuation. A common error is to subtract the $4.00 million continuation cost at time 0 in the staged case. It is paid only if management continues, at time 1, so it belongs in the state-by-state time-1 net values.

Using the Result Carefully

The flexible NPV is the value of the entire staged plan, including its initial cost. The flexibility premium is a comparison with mandatory continuation, not an amount to add to every project NPV regardless of its structure. In the one-period model, the premium is the present value of avoided losses in states where continuation value falls below continuation cost. If all states have $V_s\ge K$, flexibility has no value in this model because management continues in every state.

Valuation also depends on how risk is treated. Risk-neutral valuation weights paired with the risk-free rate form one consistent method when the model supports those weights. Alternatively, a cash-flow valuation may use real-world expected cash flows and a risk-adjusted discount rate suitable for the cash-flow risk. The optional payoff can have a different risk from the underlying project, so mechanically applying a project's WACC to every staged payoff may be inappropriate. When the risk or timing differs by stage, make the valuation assumptions explicit and use a consistent method for each cash flow.

The essential decision rule is simple: at the information date, continue only when the value of continuation exceeds its cost. But the timing of that decision is decisive. Averaging before applying the rule describes a commitment made before information arrives; applying the rule state by state values the flexibility to wait.

Check Your Understanding

Use the stated valuation assumptions in each problem. Monetary answers are in millions unless stated otherwise.

A pilot costs $0.80 million now. In one year, states with weights 0.30 and 0.70 have continuation values of $12.00 million and $5.00 million, respectively. The continuation cost is $7.00 million, and $i=10\%$. Find flexible and mandatory-continuation NPV. Assume the weights are risk-neutral. (Answer: flexible NPV = $0.5636 million; mandatory-continuation NPV = −$0.7091 million.)
A pilot has no other cost information specified, and management wants the largest $C_0$ that makes flexible NPV nonnegative. In one year, a state of weight 0.25 has continuation value $16.00 million and a state of weight 0.75 has value $10.00 million. Continuation costs $12.00 million; $i=8\%$, with risk-neutral weights. Find the maximum $C_0$. (Answer: maximum $C_0$ = $0.9259 million.)
A pilot costs $1.50 million now. At time 1, values of $15.00 million and $9.00 million occur with valuation weights 0.40 and 0.60. Continuation costs $11.00 million and $i=6\%$. Find flexible NPV, mandatory-continuation NPV, and the flexibility premium. (Answer: flexible NPV = $0.0094 million; mandatory-continuation NPV = −$1.1226 million; flexibility premium = $1.1321 million.)
A pilot costs $0.50 million and reveals a signal in one year. The underlying high and low states each have risk-neutral probability 0.50. The signal is positive with probability 0.80 in the high state and 0.30 in the low state. Continuation values are $14.00 million in the high state and $6.00 million in the low state; continuation costs $9.00 million; $i=5\%$. Management continues only after a positive signal if its conditional expected net value is positive. Find staged NPV with the signal and staged NPV if management must decide whether to continue before observing it. (Answer: with the signal, continue only after a positive signal and NPV = $0.9762 million; without the signal, continue and NPV = $0.4524 million.)

What You'll Learn