One contractor, two companies.

Two shapes, and the difference decides everything

The distinction that matters isn't the legal label. It is whether the scope was split.

The label can't carry it, and the reason is that there are three usages in circulation. A standard management reference treats joint venture and consortium as the same thing. The standard international conditions define a joint venture to take in a consortium, an association or any other unincorporated grouping, which collapses them on purpose. And the civil-law treatment separates them sharply: a consortium is not a legal entity, its members execute their own parts and carry their own profit and loss, while a joint venture is an entity with one set of books. At least two of those are live on any given job, which is why the word tells you less than the question does.

In an integrated arrangement the partners pool people and resources into one organisation that runs the work. There is one project team, one cost system set up for the job, and one programme built by that team. Internally it behaves like a single contractor, because for the purposes of doing the work it is one.

In a split-scope arrangement the partners divide the works between them — one takes the process plant, another the civils — and each executes their part with their own people, their own systems and their own supply chain. There is one contract facing the employer and two projects behind it.

The second shape reproduces every problem of the last five weeks inside a single contract. What it doesn't reproduce is any of the machinery.

What the machinery can't reach

Between two package contractors, at least there are two contracts, two Engineers and an employer standing between them. It is awkward, as week 4 showed, and it is a structure.

Between two consortium partners there is no contract at all. There is a joint venture agreement, which is an agreement between them about how they will share the work, the money and the risk. The employer isn't a party to it. The Engineer has no visibility of it. Nothing in the construction contract governs what happens when one partner delays the other, because as far as that contract is concerned nobody was delayed — the contractor was.

So the internal interface has no notice provision, no determination, no extension mechanism and no adjudicator. Whatever exists is in the JV agreement, and a JV agreement is a corporate document about liability and profit share. It isn't drafted to resolve a programme dispute, and it is rarely written by anybody who expected one.

What the employer was told

One document does exist outside the two partners, and it is far easier to see than the agreement between them. Tendering under the standard international forms, a joint venture has to submit a letter signed by every member. It does three things: each member undertakes to the employer to be liable for the performance of the whole of the contractor's obligations, the member who leads and can bind the others is named, and the separate part of the works each member will carry out is identified — if there is one.

That last qualification is the shape question, answered in writing before the contract was awarded. Where parts are listed, the arrangement was split at the point of bidding. Where the field is empty, it was pooled. The employer has held the answer since before signature, and the planner working inside the venture usually hasn't asked for it.

The first part matters in the other direction. Liability to the employer is joint and several: each member answers for all of it rather than for its own share. So the split that decides everything internally decides nothing at all about who the employer can pursue, and a partner holding a fifth of the scope carries the whole obligation if the other one fails.

One caveat, because it is the case that catches people. Where an arrangement is put together internally after award rather than declared at tender, it doesn't reach the employer at all and no joint and several liability arises from it. That is the version most likely to exist without anybody in the project team knowing it does.

The delay that doesn't exist

Follow one event through and the consequence is stark.

Partner A hands over an area three weeks late. Partner B can't start and loses three weeks. The joint venture's completion date is threatened.

There is no claim to make. The contractor delayed itself. The employer owes nothing, no extension is due, and the Engineer has no question in front of them. The three weeks are absorbed inside the JV, and where they land depends entirely on a document the project team may never have read.

Which is the reverse of the position in week 5. There, another contractor's delay reached you as an employer risk event and became an ordinary claim. Here the same physical event produces nothing at all, because the party who caused it and the party who suffered it are the same contracting party.

Two sets of books, one figure

The reporting consequence is immediate and it lands on project controls.

Each partner records cost in their own system, on their own chart of accounts, closing on their own date. Each measures progress by whatever method they use elsewhere. The figure the employer receives is a consolidation of the two, produced by somebody, on a basis that is agreed between the partners and stated nowhere in the contract.

Everything Reporting Week 25 said about reconciling two records applies, with one difference that makes it harder: there is no reconciliation meeting with an agreed outcome, because the two parties are commercial rivals whose share of the profit depends on how the work is attributed between them.

The consolidated number can be entirely correct and still be unauditable from outside, because the working underneath it is commercially sensitive between the two firms producing it.

Whose planner are you

This is worth naming because it is uncomfortable and it is unavoidable.

A planner in a split-scope consortium is employed by one partner and producing documents for the joint venture. The programme goes to the employer in the JV's name. The delay analysis that shows where three weeks were lost also shows which partner lost them.

The professional position is the same one Reporting Week 1 set out: the reliability of the number is yours, the number isn't. What changes is that the parties who will disagree about it are on the same side of the contract, and the argument has nowhere formal to go.

Which makes the record the whole of the defence. An internal interface that is logged, dated and circulated to both partners at the time is a fact. The same interface remembered six months later is a negotiating position.

System design

An interface with no contract behind it still has to be run, and what is left to run it with is the record.

Record	Produced by	Required quality	Verified against	Feeds
Internal interface register	Project controls	One line per boundary between partners, with an owner each side	The scope split	Handover evidence · internal claims
Handover evidence	Both partners	Dated and circulated at the time, not reconstructed	Site record	Where an internal delay landed
Consolidation basis	Both partners	How two cost systems and two progress methods become one figure	The JV agreement	The employer’s monthly report
Partner cut-off dates	Each partner	Both stated, because they won't be the same	Each partner’s finance calendar	Reconciliation before issue

The second row is the one that decides internal arguments. Evidence circulated to both partners on the day is a fact both of them accepted at the time; the same event described afterwards is one company’s account of it, and the other company has their own.

Practical insight

Find out which of the two shapes you are in, and ask about systems rather than about the agreement. You will get an answer this week; the agreement you may never see.

Does your project have one cost system or two? Do you report progress on one method or consolidate two? Are your colleagues reporting to one manager, or do they have their own line back to their own company? Those three answers tell you more than the legal form does, and you can get them before lunch.

If your answer is two, draw the boundary between the partners as an interface and run it exactly as you would run one between separate contracts — a register, an owner on each side, dated evidence of every handover. You won't have a contract to enforce it with. You will have the only record of what happened, and on this arrangement that is worth more than the contract would have been.

Key takeaways

• The distinction that matters is whether the scope was split, not the legal label.
• Integrated: one team, one cost system, one programme. It behaves like a single contractor because it is one.
• Split-scope: one contract facing the employer and two projects behind it.
• Between two package contractors there are two contracts and an Engineer. Between two partners there is a JV agreement the employer never sees.
• An internal delay produces no claim. The contractor delayed itself, and the loss lands where the JV agreement puts it.
• The employer's progress figure is a consolidation of two systems on a basis stated nowhere in the contract.
• It can be correct and unauditable, because the working is commercially sensitive between the parties producing it.
• With no machinery to enforce the interface, the contemporaneous record is the whole of the defence.
• Joint venture and consortium are used three incompatible ways. Ask about cost systems and reporting lines instead.
• A joint venture tendering under the standard forms files a letter naming the leader and identifying each member's separate part of the works, if there is one. That field answers the shape question.
• Liability to the employer is joint and several. Each member answers for the whole, whatever the internal split says.

Records born here. The internal interface register between partners · the consolidation basis for cost and progress · dated handover evidence at every internal boundary · the scope field from the tender undertaking, copied at the start.

What is coming next

Some contracts are written to stop all of this happening — to put the parties in one pool, align the money, and remove the machinery that turns a disagreement into a claim.

Next week: alliancing, partnering and integrated delivery, and what a planner does when the contract is designed to suppress the claim.