Who Owns the Float? Construction Schedule Float, Explained
Total float, free float and terminal float — what each means, who owns them under your contract, and how to track float burn before it becomes a claim.
Float is the only asset on a construction project that two parties can both believe they own, spend without recording, and then fight over eighteen months later in front of an arbitrator. It does not appear on a schedule of values. It is not in the general conditions budget. Nobody signs for it. And on most commercial jobs it is gone by the time anyone thinks to look — consumed one three-day slip at a time by a schedule update nobody read past the finish date.
The finish date is the wrong place to look. A project that has burned twenty-two of its twenty-five days of float is in serious trouble and still shows "on schedule" on every dashboard in the building. Float is the early-warning system, and it only works if somebody is measuring the burn rate rather than the endpoint.
The three kinds of float, and why the difference matters
Most schedulers use "float" to mean total float. There are three quantities, and they answer different questions.
Total float is how long an activity can slip before it pushes the project's contractual completion date. It is a property of a path, not of a single activity — a chain of activities shares its total float, which is why one subcontractor can quietly spend the float that three other trades were counting on.
Free float is how long an activity can slip before it delays the start of the next activity. Free float is always less than or equal to total float. It is the number that matters to the superintendent sequencing next week; total float is the number that matters to the person who will defend the claim.
Terminal float — the least discussed and often the most valuable — is the gap between the date the contractor planned to finish and the date the contract requires completion. A contractor who submits a baseline finishing thirty days early has built thirty days of terminal float. Whether that contractor gets to keep it is the entire argument.
The distinction has teeth. An owner-caused delay that consumes total float but leaves the completion date intact may or may not entitle the contractor to an extension of time — and the answer is determined not by the schedule but by a clause most people never read.
Who owns it: three positions, three very different outcomes
There is no universal rule. Ownership depends on the contract, and the three live positions produce genuinely different results from identical facts.
| Position | Typical clause language | What happens when an owner-caused delay eats 10 days of float |
|---|---|---|
| Contractor-owned | "All float in the project schedule is for the exclusive use of the Contractor." | Contractor can claim an extension of time even though completion has not moved. Float is treated as contractor property that was taken. |
| Owner-owned | Float is a resource the owner may allocate; the contractor has no exclusive claim to it. | No extension of time until float is fully exhausted. Owner delays are absorbed free of charge. |
| Project-owned / shared | "Float is not for the exclusive use or benefit of any party but is a resource of the Project." | First come, first served. Whoever consumes the float first gets the benefit; the second party inherits a schedule with no cushion. |
Long International's survey of the approaches notes that the shared or project-float position is the one that aligns with the direction of U.S. federal case law, and that it is also the position that most requires disciplined governance — because "first come, first served" rewards whichever party is paying attention. (Long International)
The standards do address it. AACE International's Recommended Practice 29R-03 on forensic schedule analysis, the Society of Construction Law's Delay and Disruption Protocol, ConsensusDocs 200 and the FIDIC forms all engage with float and concurrency. None of them overrides an express clause in your contract. That is the point: the contract wins, and most contracts are silent.
Silence is not neutral. Silence means the question gets answered retroactively, by whichever party's expert builds the more persuasive as-built analysis — which is to say, by whichever party kept better contemporaneous records.
The float burn-down: the chart nobody builds
Here is the shape of a real problem. The line below is an illustrative eight-month sequence on a job that reported "on schedule" for six consecutive owner meetings.
Read the slope, not the value. Updates 1 through 3 give up six days — unremarkable in isolation. But the burn rate between updates 4 and 6 is roughly five days a month against a remaining balance of twelve. Any operator looking at rate rather than level knows in month five that this job finishes late. The finish date does not admit it until month seven, by which point the recovery options are the expensive ones: acceleration, added shifts, or a negotiation.
Float consumption is a records problem before it is a schedule problem
The reason float disputes are so brutal is that float is consumed by events that individually looked too small to document. A three-day RFI response. A submittal returned "revise and resubmit" for a finish color. An inspection that slipped past a Friday. None of them generated a notice, because none of them moved the completion date.
Eighteen months later the claim requires proving which party consumed which days, in order — and the evidence is an email thread, a set of daily logs of uneven quality, and two schedule files whose calendars do not match.
This is the mechanism behind the industry's data problem. Autodesk and FMI's survey of more than 3,900 construction professionals estimated bad data cost the global industry $1.85 trillion in 2020, including $88.69 billion in rework — about 14% of all rework performed that year. Thirty percent of respondents said more than half their project data was bad. (Autodesk / FMI) A float claim is that statistic with a date attached.
The first time I lost a float argument, I had the better case and the worse file. The owner's scheduler had a clean monthly record of every update, every fragnet, every changed logic tie. I had eleven months of "we told them in the OAC meeting." The schedule was on my side. The record was on theirs.
What to actually do
1. Find the float clause before you sign, and price the answer. If the contract says float belongs to the project, your terminal float is not yours — build the schedule knowing an owner-caused delay is free to the owner until the cushion is gone. If the contract is silent, expect the project-float position by default and plan accordingly.
2. Baseline the float, then track it as a number with a trend. Total float on the two or three longest paths, captured at every update, plotted. One line. If it is falling faster than the calendar, escalate — regardless of what the finish date says.
3. Watch the near-critical paths, not just the critical one. A path with eight days of float is not comfortable; it is one weather event and one late submittal from becoming the driving path. Most schedules have three or four paths within fifteen days of critical, and the critical path changes hands two or three times over a job.
4. Attribute every consumption event as it happens. Not "the schedule slipped four days" but "four days, RFI 214, response received day 19 of a 10-day review window." Attribution is cheap on the day it happens and nearly impossible to reconstruct at month twenty.
5. Give notice on float consumption where the contract allows it. Many notice provisions are triggered by delay to the work, not by delay to completion. If yours is, a float-consuming event is a noticeable event — and a notice that costs you nothing preserves an argument that may later be worth the general conditions on a two-month extension.
6. Never submit a baseline with hidden float. Padded durations and artificial constraints get discovered in forensic analysis, and the discovery costs more credibility than the float was worth. If you want terminal float, show it and negotiate for it in the clause.
The measurement that changes the conversation
Float is a leading indicator sitting inside a system almost everyone treats as a reporting output. The schedule update is filed, the finish date is reported, and the one number that would have predicted the outcome — the rate at which the cushion is disappearing, and whose events are spending it — is computed monthly and then discarded.
The projects that win float arguments are not the ones with better lawyers. They are the ones where the float balance, the burn rate and the attribution were a standing agenda item from the first update, in a record built the day the event happened rather than the year the claim was filed. By the time the finish date moves, you are no longer managing the schedule. You are documenting what happened to it.
Sources
- Float ownership positions (contractor-owned, owner-owned, project/shared) and their effect on extension-of-time entitlement — Long International, "Float Ownership – Why It Matters and Approaches" — source
- Standards that address float in delay analysis (AACE RP 29R-03; SCL Delay and Disruption Protocol; ConsensusDocs 200; FIDIC) — Long International — source
- SCL Delay and Disruption Protocol, 2nd edition (the reference text for float and concurrent delay) — Society of Construction Law — source
- Definitions of total float and free float — SmartPM, "Total Float vs. Free Float in Construction Scheduling" — source
- Bad data cost the global construction industry $1.85 trillion in 2020, including $88.69B of rework (14% of all rework); 3,900+ professionals surveyed; 30% said more than half their project data is bad — Autodesk / FMI, "Harnessing the Data Advantage in Construction" — source