Float ownership and the contract
Last reviewed 9 September 20262,699 words12 min read
The clause fight, how it's usually resolved
π© In one line: Unless the contract says otherwise, float belongs to the project and goes to whoever needs it first β which is why the number that matters is not "who owns the float" but "how much float was there on the day the event happened", and only a properly maintained update can tell you that.
π€ Who this is for: M/S. You should already understand total float, the TIA method and the three concurrency tests. Juniors should read it once to understand why the narrative reports float every month.
First, let's be honest about why this page exists
Every contractor believes the float is theirs: they built the programme, they chose the sequence, the slack is their planning margin. Every employer's representative believes the float is the employer's: it's time the contractor didn't need, so the employer can use it for a variation without granting an extension. Both sides state their position as though it were obvious. Neither position is what most contracts or most tribunals apply.
The practical result is a fight that surfaces at the first delay event and runs to the final account. The planner cannot settle who is legally entitled to float. What the planner can do β and it decides most of these disputes β is show, from the accepted updates, exactly how much float existed on the affected path at each data date, and who consumed it in which order.
π¨ The standard β what "good" looks like
| Source | What it says (paraphrased) |
|---|---|
| SCL Delay and Disruption Protocol, 2nd ed. β Core Principle 8 | Unless the contract provides otherwise, an Employer Risk Event that only uses up float does not entitle the Contractor to an EOT; entitlement arises only once the float on the affected path is predicted to fall below zero. Float is, in effect, available to whoever needs it first. |
| SCL Protocol β Core Principle 9 | The programme should identify float clearly, distinguishing total float, free float and any terminal float, so the parties can see it being consumed. |
| SCL Protocol β Core Principle 13 | Where the Contractor planned to finish early and an Employer Risk Event consumed that early-completion period, compensation may still be recoverable even where no EOT is due β the time and money questions are separate. |
| FIDIC 1999 Cl 8.4 / FIDIC 2017 Cl 8.5 | EOT is due to the extent that completion "is or will be delayed". No express float clause. The test is delay to Completion, so non-critical float is absorbed without entitlement β the project-owns-float position by default. |
| FIDIC 2017 Cl 8.3 | The programme must show the sequence, logic and float for the Works; the Engineer may reject a programme that does not. |
| NEC3 Cl 31.2 and 63.3 / NEC4 Cl 31.2 and 63.5 | The Accepted Programme must show float and time risk allowances separately. Delay to Completion is measured as the movement of planned Completion, not the Completion Date β so terminal float (the gap between planned Completion and the Completion Date) is preserved for the Contractor. Time risk allowances are the Contractor's. Total float within the network is the project's. |
| AACE International RP 29R-03, Forensic Schedule Analysis | Float ownership must be established from the contract before analysis; where silent, the analyst should state the assumption applied and apply it consistently across all windows. |
| Glenlion Construction v The Guinness Trust (1987); Ascon Contracting v Alfred McAlpine (1999) | Orientation only, Check the citations in a law report before relying on them. The first: a contractor may plan to finish early but the employer is not obliged to help it do so. The second: float is not the exclusive property of either party and is consumed by whichever delay reaches it first. |
π’ Rule: read the Particular Conditions first; if they are silent, float belongs to the project on a first-come basis β so your job is to record how much there was, on which path, at every data date.
How it actually works
1. Three kinds of float, three different answers. Most of the argument is people using one word for three things.
| Term | What it is | Default ownership (silent contract) | NEC position |
|---|---|---|---|
| Total float | Slack on a non-critical path before it becomes critical | Project β first come, first served | Project |
| Terminal float | Gap between the contractor's planned completion and the contractual completion date | Contested under FIDIC; usually no EOT but possible cost | Contractor's (Cl 63.3 / 63.5) |
| Time risk allowance / contingency | Explicit duration added for the contractor's own risks, shown as such | Contractor's, if it is visible and named | Contractor's, must be shown (Cl 31.2) |
2. The first-come rule and why order matters. Suppose a path has 15 days of float. An employer delay of 10 days lands first: float drops to 5, no EOT, no LDs. A contractor delay of 8 days lands second on the same path: float goes to β3, and the contractor is 3 days culpable. Reverse the order β contractor delay first, employer delay second β and the outcome is 3 days of EOT with the contractor absorbing nothing. Same two events, same magnitudes, opposite result. This is the whole reason contractors want to own float and employers want to keep it shared, and it is why the sequence of events in the updates is the evidence.
3. What the contract actually says. Check in this order:
| Where to look | What you're looking for |
|---|---|
| Particular Conditions amending Cl 8 (FIDIC) or Z-clauses (NEC) | Express float clauses. Gulf employer-amended FIDIC commonly adds "float is not for the exclusive benefit of either party" (project owns) or, less often, "all float is for the benefit of the Employer". |
| Employer's Requirements / scheduling specification | Definitions of float, requirements to show it, sometimes a stated ownership position that the Conditions don't repeat. |
| Programme acceptance correspondence | Whether the engineer accepted a programme showing early completion or a named contingency without objection. |
| General Conditions | If nothing above applies, FIDIC is silent (project owns by operation of the "is or will be delayed" test); NEC is explicit on terminal float and TRA. |
4. Early completion under FIDIC. A contractor programme that finishes 6 weeks before the Time for Completion is legitimate. If an employer delay consumes those 6 weeks, no EOT is due because the Time for Completion has not been exceeded β but the contractor's time-related costs for the period may be claimable if the early finish was genuinely achievable and the delay is proven (SCL Core Principle 13). The practical difficulty is proof: an early-finish programme built on optimistic durations will not survive scrutiny. State the planned completion and the terminal float in the SBM and every narrative so the position is on record from day one.
5. Making float visible β what the planner controls.
- Keep a Total Float column in every submitted layout; never suppress it.
- Show the planned completion milestone and the contractual completion milestone as two separate activities with the gap between them stated in the narrative. The contractual milestone carries the Finish On or Before constraint; the planned one carries none.
- Name contingency activities as such β "TRA β FaΓ§ade procurement 10d" β never buried in a duration. Under NEC this is mandatory; under FIDIC it is the only way to argue later that the allowance was yours.
- Record float on the top three paths in every monthly narrative with the trend. When an event lands, the TIA base is the last accepted update before it, and that update is your proof of how much float existed.
- Archive the Schedule Comparison for every update so float consumption can be traced event by event.
π Contract sub-section: how it usually resolves
In practice, on FIDIC-based Gulf contracts, three outcomes cover almost every case:
| Situation | Usual outcome |
|---|---|
| Employer event on a path with float, float not exhausted | No EOT. Contractor records the consumption; if a later contractor delay tips the path negative, the contractor argues the employer used the margin β usually unsuccessfully on time, sometimes successfully on cost. |
| Employer event exhausts float and delays completion | EOT for the portion beyond zero float only. The days that were "float" before are not recovered. |
| Contractor planned early finish, employer delay eats it | No EOT. Prolongation cost arguable if the early finish is proven credible; rarely awarded in full. |
Under NEC the picture is cleaner: terminal float and TRA are protected, and the compensation event assessment moves planned Completion, so the contractor's margin survives an employer delay intact. That is one reason NEC contractors show a planned Completion earlier than the Completion Date and FIDIC contractors often don't dare to.
π₯ Where people go wrong
- Hiding float in lags and constraints. A 10-day FS lag "for coordination" or a Start On constraint that holds a start is float disguised as logic. It fails the review and it fails the ownership argument β you can't claim slack you pretended didn't exist.
- Stating "the float is ours" without a clause. Under a silent FIDIC contract that position has no basis. Say what the contract says, then say what the updates show.
- Confusing terminal float with total float. They have different owners under NEC and different treatment under SCL. A narrative that reports "6 weeks float" without saying which one is inviting the wrong argument.
- Not recording the float at the data date. If the last accepted update before an event doesn't exist or wasn't accepted, nobody can prove how much float was there. The first-come rule cannot be applied and the tribunal falls back on assumptions β rarely in the contractor's favour.
- Treating the SCL Protocol as the contract. It is guidance. It applies only where the contract is silent and even then only if the parties or tribunal choose to follow it. Cite it as the fallback, not the rule.
- Padding durations instead of declaring TRA. A concreting duration inflated by 20% is invisible contingency that the employer will consume without anyone noticing. A named TRA activity is a defensible allowance.
- Letting the reviewer pull planned completion onto the contract date. Employer reviewers sometimes demand the programme "align" with the Time for Completion, which means constraining the finish and destroying terminal float. Push back: the contract requires a programme showing how the Works will be executed, not a programme that finishes on a prescribed day.
βοΈ When you're challenged
"The float is the contractor's. You didn't need those 10 days, so we're taking them." The contract is silent on ownership, so float on the affected path is shared and goes to whoever reaches it first. Your instruction reached it first and consumed 10 of the 15 days β recorded in Update 07. If our subcontractor now slips 8 days on the same path, we'll be 3 days negative, and I'll be pointing at Update 07 when we discuss whose 3 days those are.
"Your programme finishes six weeks early. We're not bound by your early date." Agreed β you're not obliged to help us finish early. But the planned completion and the six weeks between it and the Time for Completion have been in every accepted programme since Baseline 0, and the SBM states our position. If an employer event consumes that period, we won't claim time, but we will notify and claim the time-related cost under Cl 20.
"You've buried three weeks of float in that lag." If that lag doesn't represent a physical necessity I'll remove it and show the float openly in the Total Float column. I'd rather have visible float I can defend than hidden float I can't.
"Why should the employer pay because you chose to plan aggressively?" The employer doesn't pay for our planning; it pays if its own delay event caused us to be on site longer than we otherwise would have been. Whether the early finish was achievable is a matter of evidence β durations, resources, the progress trend before the event β and that's what the analysis will test.
π Related pages
- Total float β what the number is before anyone argues about owning it.
- Time Impact Analysis β the method that proves how much float existed when.
- Concurrent Delay β the argument that follows immediately after the float argument.
- Extension of Time Basics β the entitlement test that float feeds into.
- NEC Accepted Programme β terminal float and TRA under NEC.
- FIDIC Time Clauses β Cl 8 in both editions.
- Schedule Basis Memorandum β where the contractor's float position is first written down.
- Schedule Narrative β reporting float on the top paths every month.
βοΈ Worked example
A data centre in Abu Dhabi, FIDIC 1999 based, Particular Conditions silent on float. Path: free-issue switchgear delivery β LV room fit-out β energisation β IST β completion. At Update 04 (data date 1 March 2025) the path shows 15 working days total float against the contractual completion.
Sequence as it actually happened:
| Update | Data date | Event | Party | Days | Path float after | Effect |
|---|---|---|---|---|---|---|
| 04 | 1 Mar 25 | β | β | β | +15 | Baseline position on this path |
| 06 | 1 May 25 | Switchgear delivered late | Employer | 10 | +5 | Notice issued under Cl 20.1. No EOT: completion not delayed. |
| 08 | 1 Jul 25 | LV fit-out subcontractor under-resourced | Contractor | 8 | β3 | Path now critical. Contractor culpable for 3 days. |
Same events, order reversed (hypothetical for the narrative):
| Update | Event | Party | Days | Path float after | Effect |
|---|---|---|---|---|---|
| 06 | LV fit-out under-resourced | Contractor | 8 | +7 | No effect on completion. |
| 08 | Switchgear late | Employer | 10 | β3 | EOT 3 days due. Contractor culpable for nothing. |
What the planner did that mattered: Update 04 had been formally accepted by the Engineer, so the +15 was undisputed. The TIA for the switchgear event was run on Update 05 (the last accepted update before delivery slipped), fragnet in its own WBS node, showing consumption from +15 to +5 with zero movement of completion β which is exactly why the Engineer's "no EOT" response was correct and why the contractor's later attempt to argue "the employer used our margin" failed on time. The cost claim for 10 days of standby on the LV crew was a separate, live argument under Cl 20.1, supported by the same records.
Under NEC4 on the same facts: the switchgear compensation event would have been assessed against the Accepted Programme at the time of the event; planned Completion didn't move, so no change to the Completion Date β but the contractor's terminal float of 4 weeks (planned Completion to Completion Date) would have remained untouched, and the later LV slip of 8 days would have consumed 3 days of that terminal float rather than putting the contractor into delay damages.
π References
- Society of Construction Law, Delay and Disruption Protocol, 2nd ed. (2017) β Core Principles 8, 9, 13; Part B guidance on float (check against your copy).
- FIDIC Conditions of Contract for Construction, 1999 β Cl 8.3, 8.4, 20.1 (check the edition in your contract).
- FIDIC Conditions of Contract for Construction, 2017 β Cl 8.3, 8.5, 20.2 (check the edition in your contract).
- NEC3 ECC β Cl 31.2, 63.3; NEC4 ECC β Cl 31.2, 63.5 (check the edition in your contract).
- AACE International RP 29R-03, Forensic Schedule Analysis β sections on float ownership (check against your copy).
General guidance, not legal advice. Contract wording, editions and amendments differ on every project. Read your own contract and take professional advice before relying on anything here in a claim or a dispute.
From the field
Experience from working planners. Unreviewed β read it as experience, not guidance.
Add what you know about float ownership and the contract. What worked, what the consultant pushed back on, what you would do differently next time. A paragraph is plenty.
Contributors get their name and one link on the site β your own templates, course or consultancy. We take nothing and hold nothing.
Add your experience