JV and consortium governance in construction: participation, voting, risk and control before the bid is signed

The first governance mistake is to negotiate the participation percentage and postpone everything else. A 60/40 or 40/30/30 split does not by itself answer who controls the programme, who appoints the Project Director, who can sign a variation, who funds a cash shortfall, who decides whether to settle a claim, or what happens if one partner fails to mobilise. Participation, economic exposure and decision rights should be designed together. If they are separated unintentionally, the party carrying most of the capital or liability may discover that it does not control the decisions creating that exposure.

JV and consortium labels should not be used as shortcuts. ICC's construction models distinguish them mainly by internal performance-risk sharing: a JV typically shares broader project risks, rights, benefits and profits in proportion to participation, while a consortium can preserve stronger internal responsibility for each member's own scope even though the employer may still require joint-and-several liability. In MDB procurement, the employer-facing test may be simpler: all JV members can be required to be jointly and severally liable, with a designated leader authorised to bind them. The internal agreement therefore has to manage recourse between partners even when the employer can pursue any one of them.

A workable governance architecture normally separates ordinary operational decisions from reserved matters. Routine decisions may be delegated to the Project Manager or approved by simple majority. Reserved matters should capture decisions capable of changing the partners' economic or legal exposure: bid price and major discount, material scope change, settlement of a significant claim, guarantees, borrowing, related-party subcontracting, replacement of key personnel, admission or removal of a member, major litigation or arbitration, termination, insolvency response and amendments to the JV agreement. The threshold can be unanimity, supermajority or weighted voting, but it should be chosen deliberately rather than inherited from the number of committee seats.

The governance design must also survive stress. During tendering every partner can appear aligned; after award, losses, delayed payments, LD exposure, procurement overruns and financing calls create different incentives. The agreement should therefore specify cash-call timing, default interest, dilution or funding remedies where legally appropriate, step-in rights, replacement of a defaulting member, access to the defaulting member's design and documents, continuation of guarantees, treatment of profit/loss, and whether the non-defaulting members can complete the defaulting scope and recover the cost. A governance model is only proven when it explains how the project continues after one member stops cooperating.

Gokbilge's role in this type of structure is technical-commercial governance rather than legal drafting. Before bid submission, the project can be mapped through a participation matrix, scope matrix, RACI, voting/reserved-matter matrix, guarantee-capacity matrix, cash-flow contribution schedule and interface register. The objective is to ensure that the percentage written next to each partner is consistent with what that partner must finance, deliver, guarantee, control and absorb if another member fails.

JV and consortium structures are highly project-, jurisdiction- and contract-specific. The bidding document, employer contract, JV/consortium agreement, corporate approvals, competition rules, tax treatment and applicable law govern the live arrangement. This series explains engineering, tender, commercial and governance risks; it is not legal, tax or accounting advice.