The No-Project Case:
Pricing the Decision
to Do Nothing

The decision to do nothing is the only capital decision most organizations make without a business case. A project must survive hurdle rates, sensitivity reviews, and signatures. The alternative it is measured against, waiting, enters the comparison priced at zero and renews itself every quarter nobody examines it.

Section 01The default that is never audited

When a power project reaches an investment committee, it arrives carrying a model: capital cost, operating cost, schedule, risks, and the name of a person accountable for each. The alternative it competes against arrives carrying nothing. No sponsor, no model, no signature, because in nearly every governance system inaction requires none. The project is audited; the default is not. Ties go to the unexamined option, which is the one outcome a decision process should never produce.

The remedy is to treat the no-project case as what it actually is: a full path with cash flows on both sides of the ledger. On the cost side it books the margin on load the site cannot serve, the drift of the tariff the site stays on, the escalation of the project being deferred, and the market windows that close on their own schedules. On the credit side it books preserved capital, avoided operating burden, options kept alive, and the information that arrives free with the passage of time. None of these lines is exotic. They are simply never assembled in one document, because nobody is paid to assemble them. Every seller's economics improve when the do-nothing baseline stays vague, and most owners' processes let it.

In the eight-path framework that opened this series, no-project is path eight, and it is mandatory in every comparison, because it is the baseline every other path must beat. This paper is the mechanics of that row: what belongs in it, what honestly offsets it, how to put it on a monthly meter, and how to tell a deliberate wait from an expensive drift.

Section 02The four cost lines of doing nothing

Four lines cover nearly everything that belongs on the cost side. Each has a defensible basis, and each has a characteristic way of being miscounted.

Line one: deferred margin

If a capacity constraint is blocking production, tenancy, or service, the cost of a month of delay begins with the contribution margin of whatever that month would have produced. Margin, not revenue: the revenue figure flatters the case for action, and a baseline built on flattery gets discounted the first time a board member reads it closely. The discipline runs the other way as well. Count only demand that is evidenced: contracted, committed, or probable enough that the sales organization will put its name to a percentage. Growth that exists only in a strategy deck enters this line at a stated probability or not at all.

The most common failure here is not overstatement. It is omission. The site loses nothing visible today, so the line is left blank. Deferred margin is invisible precisely because the capacity that would have earned it was never energized. The absence of a variance report is not the absence of a cost.

Line two: tariff drift

While the decision waits, the site keeps buying every kilowatt-hour at the applicable tariff, and in California that standstill has a documented price level and a documented trend.

Federal data put California's average industrial retail price at 20.20 cents per kilowatt-hour in May 2026, against a national industrial average of 8.71 cents.¹ The trend has not been flat: the Public Advocates Office at the California Public Utilities Commission reports ten-year increases in residential average rates of 69, 98, and 101 percent across the state's three large investor-owned utilities between January 2016 and March 2026, well ahead of general inflation over the same report's comparison window.²

Two cautions keep this line honest. First, the record is not a forecast. The same report shows three-year changes of 6 to 9 percent, a materially slower pace than the decade average, and rate design for large customers is its own subject. Carry escalation as a sourced range with an as-of date, never as a single blessed number. Second, this line counts only as a difference against the path being compared. If the candidate path is expected to deliver energy below the tariff, every month of waiting pays that spread. If the candidate path prices above the tariff, waiting collects the spread, and the model must be allowed to say so. Tariff drift is a signed quantity, not a scare figure.

Line three: escalation on the deferred project

The project you defer is rarely the project you later buy. Prices move and lead times move, and they do not move together or in one direction. Over the last decade some equipment classes have posted sustained price declines while others, particularly long-lead electrical gear, have risen sharply in tight markets, and a deferred order re-enters the line behind every order placed in the interval. The defensible treatment is procedural rather than predictive: date every quote, record its validity window, carry escalation as a range sourced at pricing time, and let the range be asymmetric where the market plainly is. A no-project case that shows zero on this line is asserting that markets hold still as a courtesy to the undecided. They do not.

Line four: windows

The first three lines accrue smoothly. The fourth does not. Leases expire on dates. Land options lapse on dates. Entitlements and permits carry validity periods. Customer commitments have outside dates with remedies attached. Statutes qualify property by dates: under current federal law the investment tax credit for qualifying clean-energy property stands at 30 percent, with statutory adders that must be individually qualified rather than assumed, and both the rate and the qualification tests carry effective dates that a sufficiently delayed project can cross.⁵ Confirm any incentive line with qualified tax counsel, and stamp it with the date it was checked.

Windows are the least modelable line and often the most decisive, because they convert smooth cost curves into cliffs. The correct presentation is a dated list, not a percentage. A baseline that blends a month-30 lease expiry into an annual carrying cost has hidden the one fact the committee most needed to see.

Section 03What waiting buys

A no-project baseline built only of costs is a sales document pointed the other way. Waiting has real value, and the framework earns the right to be believed by counting that value as carefully as it counts the costs.

The formal version comes from investment theory. When an investment is irreversible and the environment is uncertain, committing capital extinguishes the option to commit later with better information, and that option has value a bare net-present-value test ignores. Dixit and Pindyck built the standard treatment of this logic three decades ago, and its practical teaching survives translation into committee English: a project should beat not only zero but also the value of the waiting it forecloses.⁴ Experienced boards intuit this, which is why waiting feels safe. The job of the baseline is to make the intuition pay rent.

An option to wait is worth the most when three conditions hold together. The uncertainty is material. It resolves on a knowable date: a rate proceeding concludes, a utility study returns, a lease negotiation lands, interval data from existing operations accumulates into a real load shape. And the running cost of delay from Section 02 is low. Strike any one of the three and the value collapses. Uncertainty that never resolves is not an option; it is weather. A resolution date paired with a high monthly delay cost is an option the site cannot afford to hold. The credit side of the baseline should therefore be written as a list of named information events, each with an arrival date and the decision it would change. An unnamed flexibility credit is the mirror image of an unpriced delay cost, and it deserves the same treatment: strike it.

Three further credits are legitimate and often decisive. Waiting preserves capital and debt capacity for the rest of the business, which may have better uses for both; an energy project competes for the balance sheet, not just against the tariff. Waiting avoids an operating commitment, because any self-supply path is a plant with staffing, maintenance, and fuel exposure attached, and an organization can rationally pay a documented premium not to run one. And waiting hedges the owner's own forecast: when the demand that justifies the project is itself the largest uncertainty, the no-project case is partly protection against building for a load that never arrives. Each of these belongs in the baseline in plain words even where it resists a number.

A waiting position with a named information event and a dated review is a strategy. Without them it is drift, wearing the costume of prudence.

Section 04Constructing the baseline

The mechanics are not complicated. What they require is the same documentation standard the rest of a study carries: sourced inputs, as-of dates, and estimates labeled as estimates.

1. Fix the clock. A stated start date and a horizon that reaches the furthest dated commitment in view, so no cliff sits conveniently out of frame.
2. Put the continuous lines on a monthly meter. Deferred margin, tariff drift as a signed difference against each candidate path, and escalation accrual, each with its source and date.
3. List the cliffs as dated events. Window items are never annualized; they appear on a timeline with their documents attached.
4. Write the credit side as named events. Every option credit is tied to a specific information arrival, the date it lands, the decision it would change, and the date it expires.
5. State the result twice. A cost per month of waiting, and a dated cliff list. The pair, together, is the no-project baseline.

A worked example makes the shape concrete. Every figure in the table below is illustrative, chosen for arithmetic clarity; none is drawn from an actual site, an actual tariff filing, or any study. The hypothetical is a continuous-process manufacturer whose expansion is blocked pending additional electric capacity.

On these illustrative inputs, the site pays roughly $242,000 for each month of waiting. Few organizations would approve a change order of that size without a signature; this one renews itself monthly without one. The arithmetic is deliberately unremarkable. What changes decisions is not the sophistication of the model but the existence of the number, because the moment waiting carries a monthly price it has to compete like everything else. And the comparison cuts both ways. On different inputs, with a thinner deferred margin, a stronger information event, and a candidate path priced above the tariff, the same table can favor waiting decisively. The table does not know which answer it is supposed to produce. That is exactly what makes it worth building.

Section 05Waiting as a position, not a drift

When a committee chooses to wait, the choice should be dressed like the decision it is: an owner, a set of written triggers, and a dated review. The triggers are ordinary. The utility's written service date moves beyond a stated tolerance. A rate proceeding concludes. An equipment quote's validity lapses. The lease milestone enters its notice window. Measured load crosses a stated threshold. A statute or program the project relies on changes status. Any trigger firing reconvenes the decision, and every review re-prices the no-project case with fresh as-of dates, at the same standard as every other path.

One waiting position deserves particular suspicion: the wait pegged to a general sense that the grid will catch up. The machinery that adds supply to the grid is itself queued. Berkeley Lab's national study of interconnection queues found that the median duration from request to commercial operation for new generation doubled, from under two years for projects built between 2000 and 2007 to more than four years for those built between 2018 and 2024, and that only 13 percent of the capacity requesting interconnection between 2000 and 2019 had reached commercial operation by the end of 2024.³ Those are generator queues, not load requests, and none of it means any particular utility date will slip. It means something narrower and more useful: an undated wait leans on a system whose own timelines have lengthened, and is therefore not a plan. If the position depends on the grid, the trigger must be a written, dated utility statement, and each review must re-test it.

Section 06The row that pays the analyst nothing

A defensible no-project section, in any study by any author, contains six things:

• The four cost lines, each sourced and dated or labeled illustrative, with the signed tariff-drift convention stated rather than implied.
• The credit side as named information events with arrival dates, plus the qualitative credits in plain words.
• The result stated twice: a monthly figure and a dated cliff list.
• The triggers, the review date, and the name of the person who owns the waiting position.
• The conditions under which waiting wins, stated as plainly as the conditions under which it loses.
• An as-of date on every input, without exception.

The no-project row earns a technology seller exactly nothing, which is why it is so often missing from documents that call themselves studies, and why its presence, priced and dated, is the fastest integrity test an owner can run on any analysis put in front of them. The test binds its authors too. Our fee is fixed and identical whichever answer a study reaches, and when the honest answer is wait, the study's job is to say wait, with the price attached and the review date set.

Doing nothing is a path. It has costs that compound quietly and credits that are real but expire. Price it like everything else, put a date on its next review, and the most common failure in capital energy decisions, the unexamined default, disappears from the process. What remains is a choice among priced alternatives, which is all a good decision ever was.