Real Options Valuation: Binomial Tree Method With Worked Examples

A real option is the right (but not obligation) to take an action on a real asset — expand a factory, defer a project, abandon a money-losing operation, switch inputs in production. The action becomes valuable when uncertainty resolves favorably. Standard NPV ignores this flexibility because it assumes a fixed plan. Real options valuation captures the value of being able to adapt. The binomial tree method models the underlying asset value as moving up or down each period and computes option value by working backward from the terminal nodes. Real options matter most when (1) there is significant underlying uncertainty, (2) the option is exercisable at meaningful cost, and (3) management can credibly take the action when conditions warrant. For the four core types: option to defer (delay a project until uncertainty resolves), option to expand (scale up if demand materializes), option to abandon (exit if losses mount), and option to switch (change inputs/outputs as relative prices change).

Standard NPV computes the discounted cash flows of a planned course of action. If you can adapt — abandon a failing project, expand a successful one, defer until you have more information — those choices produce additional value not captured in the static NPV. Example: a project has standard NPV = -$5M. Standard rule: reject. But if there's an option to abandon the project after 1 year (recovering $30M of capital if it's going badly), then in the bad scenario you don't lose the full projected losses — you cut and run. The expected value of the project including the abandonment option could be POSITIVE even when standard NPV is negative. Research and oil/gas companies routinely use real options because their projects have huge flexibility: drill, observe results, drill more if good, abandon if bad. A pharmaceutical company developing a drug has option to abandon at each clinical trial phase — the project's true value can't be calculated as a single deterministic NPV. The more uncertainty there is in the underlying value, the more valuable the embedded option becomes. Volatility is not a cost in option valuation — it's a benefit (Black-Scholes intuition: higher volatility = higher option value).

Build a tree where the underlying asset value moves up by factor u or down by factor d each period. u = e^(σ√Δt) where σ is the annual volatility and Δt is the period length in years. d = 1/u (typical Cox-Ross-Rubinstein convention). Risk-neutral probability of an up move: p = (e^(rΔt) − d) / (u − d), where r is the risk-free rate. Probability of a down move: 1 − p. To value the option: at each terminal node, compute the option payoff (max of intrinsic value or zero, like a call option). Then work backward, discounting expected payoffs at the risk-free rate using the risk-neutral probabilities. Option value at any node = (p × Up node value + (1−p) × Down node value) × e^(−rΔt). For American options (exercisable any time), at each node compare the rolled-back continuation value to the immediate exercise value, take the maximum. This is the standard mechanics. The trick in real options is mapping the financial-option intuition to the real-asset decision: what is the underlying asset, what is the strike, what is the time horizon, what is the volatility?

A company is considering building a new factory. The factory costs $100M today (or $100M in 1 year if deferred — fixed cost), and will produce $115M of NPV-equivalent cash flows. Volatility of the factory's value: 30% annually. Risk-free rate: 4%. **Standard NPV:** $115M − $100M = $15M. Build today. **Real option to defer 1 year:** wait 1 year, observe value, then build only if value exceeds $100M. Binomial tree (1 step): - u = e^(0.30 × √1) = e^0.30 = 1.350 - d = 1/1.350 = 0.741 - Up value: $115M × 1.350 = $155M - Down value: $115M × 0.741 = $85M - p = (e^0.04 − 0.741) / (1.350 − 0.741) = (1.041 − 0.741) / 0.609 = 0.493 If up: NPV = $155M − $100M = $55M. Build. If down: NPV = $85M − $100M = -$15M. Don't build. Payoff = $0. Option value today = (0.493 × $55M + 0.507 × $0) × e^(-0.04) = $27.1M × 0.961 = $26.0M. Value of deferral: $26M − $15M (build today value) = $11M. Worth waiting. The option to defer captures $11M of additional value. Standard NPV said 'build today $15M'; with deferral option, the project is worth $26M.

A company is considering a $200M factory. The standard NPV is borderline: $20M positive. The kicker: if the factory succeeds, the company can build a second $200M factory worth $300M, a $100M gain it takes only in that case. Treat the decision as a one-period tree with a 50% risk-neutral probability of success and a 4% risk-free rate: - Up state (success): build the second factory, adding $300M − $200M = $100M - Down state: do not expand, adding $0 Expansion option value = (0.50 × $100M + 0.50 × $0) / 1.04 = $48.1M. Total project value with expansion option = original NPV + expansion option = $20M + $48.1M = $68.1M. Without the expansion option modeling, this project might be approved at $20M NPV, but its economic value is about $68M, more than 3x as much. The first factory is partly an option on the second factory.

A pharmaceutical company has a Phase II drug trial. Cost to complete trial: $50M. If the trial succeeds (40% probability), the company can fund Phase III ($100M additional cost, 70% success probability); a drug that passes Phase III is worth $200M. If Phase II fails, terminate (no further investment). Time value is left out to keep the arithmetic visible. Committed to every stage (Phase III funded whatever Phase II shows): NPV = −$50M − $100M + 0.40 × 0.70 × $200M = −$150M + $56M = −$94M. With the embedded abandonment option (Phase III funded only after a Phase II success): - 40% Phase II success → continue to Phase III: 70% × $200M − $100M = $40M expected. Continue (positive NPV). - 60% Phase II fail → abandon, save the $100M Phase III cost. NPV with abandonment = −$50M + 0.40 × $40M = −$50M + $16M = −$34M. Still negative, so the project is rejected either way, but the option is worth −$34M − (−$94M) = $60M: the Phase III cost saved in the 60% of cases where Phase II fails (0.60 × $100M). Real projects almost always have built-in stage gates. The error in capital budgeting is committing to all stages upfront and computing 'expected NPV' as if cash flows are forced. Modeling stages as separate options (each exercisable based on prior success) prevents this overestimation of risk.

**Pitfall 1: Applying real options when there isn't real flexibility.** A 'real option' to expand only matters if the company can actually credibly exercise it. If financing or organizational constraints prevent exercise, the option has no value. Be honest about exercisability. **Pitfall 2: Assuming volatility from financial markets without justification.** Stock volatility doesn't always translate to underlying real-asset volatility. For commodities (oil, copper) the link is reasonable. For idiosyncratic projects, you're guessing. **Pitfall 3: Double-counting value.** The standard NPV already prices in expected value of probabilistic outcomes. Adding 'option value' on top is double-counting unless you carefully separate: standard NPV under forced commitment, vs option-augmented NPV under flexibility. The DIFFERENCE is the option value. **Pitfall 4: Using complex methods when simple decision trees would suffice.** For projects with discrete decision points and well-defined probabilities, a decision tree is more transparent than a binomial tree. Reserve binomial methods for continuous-volatility cases. **Pitfall 5: Ignoring competitive dynamics.** A real option is more valuable if you have monopoly access. If competitors have the same option, your value is reduced because exercise drives down the underlying. Real options work cleanest for proprietary projects (patents, exclusive licenses, unique resource access). This content is for educational purposes only and does not constitute financial advice.