One contract asks who pays. The other asks what now.

One contract asks who pays. The other asks what now

Eighteen weeks in one family of contracts is long enough. It is worth stepping outside it, because the comparison shows something the inside view can't: what this contract is actually built to do.

Everything in this track has been retrospective. An event happens, and then a machine starts — a notice within twenty-eight days, records made at the time, a claim with a contractual basis, a determination, a period to disagree. Every one of those steps looks backwards at something that has already occurred, and asks who carries it.

The NEC family starts from a different question. Not who pays for what happened, but what do we do now. And the machinery it builds around that question changes the job of everybody administering the contract, including yours.

FIDIC already has the idea

Start with what the contract in front of you already contains, because it's more than most people realise.

Week 9 opened sub-clause 8.4. Each party has to warn the other, and the Engineer, about known or probable future events that might affect the workforce, the finished works, the price or the programme. The Engineer owes the same warning to both parties. And the Engineer may respond by asking for a proposal — the request-for-proposal route from Week 7, arriving before the damage instead of after it.

That is an early warning system. It is mutual, it covers cost and time, and it sits in a contract most people describe as adversarial.

Now ask what happens if nobody does it. Nothing. There is no register, no meeting, no consequence for staying quiet, and no link between the warning and what you can later recover. A duty with no consequence attached behaves like a suggestion, and on most FIDIC jobs that is exactly how it behaves.

What NEC does with the same idea

The NEC approach takes that clause and makes it load-bearing.

Early warning becomes a standing process rather than a letter: a register both parties maintain, and meetings held to work out what to do about the items on it. The point isn't the document. It is that the conversation happens while there are still options — a different sequence, a different supplier, a scope decision somebody can still take.

Change is handled as a compensation event, and the difference from a variation is the direction of travel. Under FIDIC, work happens and the value is argued afterwards; Week 7 showed how much of that argument turns on rates and records assembled later. Under NEC the assessment is made prospectively, from a forecast, and once agreed it is done. You find out what the change costs while you still have the option of not making it.

And the programme carries real weight. It is the tool the assessment is made with rather than a document submitted monthly and filed.

The part that is uncomfortable

It would be easy to present this as the better contract. That isn't what the comparison shows.

Prospective assessment means agreeing a number before the work is done, which means somebody is wrong. A forecast that turns out generous stays generous, and one that turns out thin stays thin. The certainty is bought by giving up the right to correct it later.

It also demands more, continuously, from both organisations. Early warning meetings need people with authority in the room every month, not just when something breaks. An employer who staffs the job thinly gets the paperwork without the process, and then has all the obligations of a collaborative contract with none of the benefits.

And the mechanism is only as good as the forecast behind it. Risk Week 9 spent an article on the fact that the shape decides the answer. A prospective assessment is a forecast with a contract wrapped around it, and everything Track 3 said about optimism and single-point estimates applies with the money already committed.

What actually changes for you

The practical difference for project controls is bigger than the legal one.

Under FIDIC the planner supports the commercial case. The programme evidences a delay that has happened; the records prove an event that occurred; the analysis comes after the fact and often long after. That is the world Week 8 described, where three quarters of a claim is built while the work is happening and the fourth quarter is assembled at the end.

Under an NEC-style contract the planner is in the transaction. The programme is what the change is priced from, so it has to be current, logical and defensible in a way that a monthly submission never has to be. A schedule that's three weeks stale isn't an administrative failing; it's a commercial exposure.

Which means the discipline from Schedule Week 17 — updated perfectly and completely wrong — stops being a quality argument and becomes a money argument.

Using the collaborative half of the contract you have

Most readers won't get to choose the form. So the useful question is what transfers.

Sub-clause 8.4 is already there, and it is mutual. Nothing stops you running an early warning register on a FIDIC job, circulating it, and asking for it as a standing item at the monthly meeting. It has no contractual teeth, and it doesn't need any — the value is the conversation happening at all.

The request-for-proposal route is already there too. Week 7 noted how rarely anybody suggests it, and that a surprising number of engineers say yes when asked. That is prospective assessment, available inside the contract you are already on, for the price of one sentence.

And the DAAB from Week 18 is the same instinct again: a standing board, informal assistance, a conversation before there is a dispute. The 2017 editions moved a long way in this direction. Most projects simply haven't moved with them.

Practical insight

Start an early warning register this month, whatever contract you are on. One line per item: what might happen, who raised it, what it could affect, and what was decided. Circulate it before the monthly meeting rather than at it.

Watch what the register does over two or three months. Some items will resolve themselves, which is fine — that's the process working. What matters is the ones where somebody makes a decision earlier than they otherwise would have.

Ask for a proposal on the next change of any size. One sentence, in writing, asking whether the Engineer would like a price and time effect before instructing. Under FIDIC that is 13.3.2; under NEC it is the default.

And keep your programme in the condition a prospective assessment would require, even where nobody's going to make one. Logic that calculates, no hard constraints propping up the dates, an update that reflects last week rather than last month. If you ever move to a collaborative form, that is the standard on day one. If you never do, you've got a better programme.

Key takeaways

✔ FIDIC is retrospective by design: an event, a notice, records, a claim, a determination. NEC asks what to do next instead of who pays for what happened.

✔ Sub-clause 8.4 already requires mutual advance warning of probable future events affecting people, works, price or programme.

✔ Nothing follows from ignoring it, which is why a clause that exists in every FIDIC contract is used on almost none of them.

✔ NEC makes the same idea load-bearing: a maintained register, meetings while options still exist, and change assessed prospectively from a forecast.

✔ Prospective assessment isn't free. The number is agreed before the work, so a forecast that turns out wrong stays wrong.

✔ The programme stops being a submission and becomes the pricing tool, which puts the planner inside the commercial transaction.

✔ Both mechanisms are available inside FIDIC already — an early warning register you choose to run, and a proposal you choose to ask for.

What's coming next

Twenty weeks of clauses, periods and mechanisms, and one thing has never been assembled in one place: the calendar. Every obligation in this track happens on a date or inside a window, and they arrive in a predictable order across a job — some in the first month, some monthly forever, some only at handover. Next week closes the track by building that year: what a contract administrator actually does in week one, what recurs, what falls due at completion, and which of it can be diarised so that none of the clocks from Week 10 ever run out unnoticed.