Real options

A DCF discounts one forecast under one fixed operating plan. Management is not bound by that plan: when the world turns out badly a project can be stopped, and when it turns out well a second tranche can be funded. Those are rights rather than obligations, which is exactly what an option is, and the standard model contains none of them.

The value comes from asymmetry. An option keeps the upside and truncates the downside, so its worth rises with the spread of possible outcomes. That reverses the usual relationship in valuation, where more uncertainty means a higher discount rate and a lower number, and the reversal is the reliable way to spot genuine optionality rather than a hopeful story.

Three forms come up in practice: the option to abandon, which covers mothballing a mine or handing back a licence; the option to expand, which covers rollouts built in tranches and capacity rights not yet exercised; and the option to delay, which is most of what a permitted but undeveloped land bank consists of.

The honest caveat matters as much as the concept. Formal valuation using Black Scholes or a binomial lattice needs a volatility of the underlying asset value, an exercise cost and an option life, and none of those is observable for a real asset. So practitioners very rarely put an option value in a client document. They use a probability weighted decision tree, they argue for the top of a DCF range, or they price the disagreement into the deal structure through an earnout or a contingent value right.

The last discipline is to test that the right is genuine. Take or pay contracts, committed work programmes attached to a licence, lender consent requirements and the political cost of closing a large employer all remove flexibility that an analysis may have assumed was there.