A fixed-price contract does not remove your exposure. It converts it into a dispute you will have later, with less information and less control.
Two large construction programmes, both in the United Kingdom, both watched closely by everyone who commissions capital work.
On one, the client entered a legally binding agreement with its key suppliers under which — in the account given by Business Insurance in October 2007 — it "accepts that it carries all of the risk for the construction project". It backed that with a project insurance policy and an incentive fund designed to balance the successes and failures of individual contractors. The agreement recorded that the client had highlighted two conditions that undermine progress: "cultural confusion and the reluctance to acknowledge risk". It also recorded that "traditional arrangements can result in a highly unproductive culture of blame and risk".
On the other, the contract was fixed price, placing the contractor at higher risk. The interviewees in that briefing associate the arrangement with a culture of blame, and the programme ended in litigation between client and supplier.
The instinct in most boardrooms is that the second structure is the prudent one. It caps exposure, it makes someone else accountable, and it converts an uncertain cost into a known one. That instinct is correct about liability and wrong about risk, and the difference is where a great deal of enterprise value is quietly destroyed.
The Strategic Context
Risk allocation is one of the few decisions in a capital programme that is genuinely irreversible. Scope can be changed, schedules resequenced, teams replaced. The commercial structure is set once, at contract award, by people who will not be present when its consequences arrive.
It is also one of the few decisions routinely made without an executive present. Procurement runs the process, legal drafts the terms, and the board sees a recommendation expressing the outcome as a price. The question the board is answering — shall we accept this price? — is not the question the structure has actually posed, which is what happens to this exposure if the party we have assigned it to cannot absorb it?
The distinction the ERANORTH collection has already drawn between risk as two-sided deviation and risk as downside is a different argument, treated in [Related article: Risk Is Not the Chance That Things Go Wrong]. This article assumes it and asks the next question: given an exposure, who should carry it, and what makes that assignment real rather than nominal?
What Leaders Commonly Misread
The first misreading is that transferring liability transfers exposure. It does not. A contract determines who is obliged to pay. Whether the assigned party can pay, and whether extracting payment costs more than the loss, are separate questions decided by that party's balance sheet and by the courts. A risk assigned to a supplier whose margin on the work is four per cent, and whose exposure under the clause is thirty, has not been transferred. It has been converted into a probability that the supplier disputes, delays or fails.
The second misreading is that the party best placed to control a risk should always carry it. This is a good principle with a limit that is rarely stated. It holds where the party can both control the risk and absorb it. Where those two conditions separate — the contractor controls the ground conditions but cannot absorb a hundred-million-pound overrun — the principle produces an allocation that looks rational and behaves badly.
The third misreading is that a hard bargain is a good bargain. One of the interviewees in the Business Insurance briefing puts the counter-case in a single line: "Designing a commercial [contract] that allows one party to win over another just gets you into court." A contract is not a competition. It is a set of incentives that will operate for several years on people whose cooperation you require.
A fourth misreading concerns what the client gives up. Transferring risk is frequently accompanied by transferring attention. The logic feels sound — if they carry it, they can manage it — and it hollows out the client's own capability at exactly the point where informed oversight matters most.
Reframing the Issue
The useful reframing is to treat risk allocation as an investment decision about where to place capability, not as a commercial negotiation about where to place liability.
Every allocation implies a division of labour. Whoever carries a risk must be resourced to manage it: to see it coming, to price it, to hold reserves against it, and to act early. If the assignment says one thing and the resourcing says another, the assignment is decorative.
That reframing produces a test with three parts, and a party must satisfy all three to carry a risk genuinely. Can they see it? — do they have visibility of the conditions that would trigger it. Can they influence it? — can their decisions change its probability or size. Can they absorb it? — would the loss be survivable at the scale the contract contemplates.
A party that fails any one of the three is carrying the risk on paper. Where all three fail, the contract has created a future dispute and priced it as a saving.
Note what this is not. It is not an argument about which constraints bind a particular initiative, which is a separate reading with its own method in [Related article: Read the Constraints, Not the Deliverables]. Constraints describe what cannot move. This concerns who pays when something does.
Strategic Analysis
What the client-retention model actually buys
The Terminal 5 arrangement is worth examining as a design rather than as a success story, because the evidence for the latter is weaker than it is usually presented.
As a design, three elements interlock. The client retains the risk, which removes the contractors' incentive to price contingency into every line and to defend it later. An incentive fund rewards performance across the supplier set rather than pitting members of it against one another. And an insurance policy converts a portion of the retained exposure into a known cost.
None of that works without the fourth element, and the source is honest that this one is an observation rather than a fact: a broker interviewed for the briefing says the client "appears to invest strongly in a lot of expertise in order to manage and be heavily involved in each of the individual subprojects", with the partnership model "underpinned by their great involvement and continuing ownership of the project", including recruiting more project managers on the client side.
That is the trade. A client that retains risk must buy the capability to manage it, and the cost of that capability is the real price of the model. An organisation that adopts the contract structure without the staffing has taken on the exposure and kept none of the control.
A caution about the most quoted number. The same briefing reports that under a traditional approach the terminal "would have been two years late and 40% over budget", attributed to the client's own commercial director in an interview published in a construction consultancy's newsletter. That is an interested party's estimate of an outcome that never happened, reported second-hand. [FACT CHECK REQUIRED] It is repeated widely and it cannot be used as evidence for anything, and an executive team that finds it in a business case should say so.
The failure mode of transfer, in two sectors
The mechanism is not peculiar to construction.
A pharmaceutical manufacturer outsources a technology transfer to a contract manufacturing organisation on a fixed-price basis, transferring the risk of yield shortfall. The contract manufacturer controls the process and can see the problem coming — two of three tests passed. It cannot absorb a failure that would consume several years of its margin on the account. When yields disappoint, it does not absorb the loss; it argues that the transferred process was not as specified. The manufacturer now has a supply problem, a dispute, and no second source, and the saving it booked at contract award was never real.
The same shape appears in defence procurement, in outsourced IT transformations, and in facilities contracts with heavy performance regimes. In each case the transfer was accepted because the assigned party could see and influence the risk. In each case it failed the third test.
The commercial-document quality that determines how much late change a contract will generate is a different mechanism with its own economics, examined in [Related article: The Front End Owns the Outcome]. That article concerns the cost of changing your mind late. This one concerns who pays when something goes wrong that nobody chose.
Decision Framework
Five steps, applicable before contract award and worth revisiting at any major variation.
1. List the material risks and name the assigned party for each. Not the clause — the party. Most organisations have this only implicitly, distributed across a contract nobody has read end to end.
2. Apply the three-part test. For each assignment, ask whether the party can see it, influence it, and absorb it. Record the answers. Any risk failing the absorb test is a candidate for retention, insurance, or a cap.
3. Price the retained capability. For every risk you keep, state what you will resource to manage it — the people, the information, the reserve. A retained risk with no attached capability is worse than a transferred one, because you have neither the control nor the counterparty.
4. Test the incentive geometry. Ask what each party does when the programme goes badly. If the honest answer is that they begin building a claim, the structure will produce a dispute regardless of anyone's intentions. Incentives that reward collective performance change that answer; incentives that reward individual performance against a fixed price do not.
5. Decide who owns the relationship after signature. Contracts are managed by people. Name the individual on your side who is accountable for the commercial structure working, and give them standing — otherwise the structure will be administered by whoever has capacity.
A supporting convention: treat every transferred risk as a credit exposure, and review it as your treasury function would review any other counterparty concentration. That framing gets the absorb test asked automatically.
From Strategy to Execution
Immediate. Take the largest capital or outsourcing contract currently in force and run steps one and two. This is a half-day exercise with procurement, legal and the delivery lead in one room, and it typically identifies two or three assignments that fail the absorb test. Those are the disputes you will have in eighteen months.
Medium term. Change what the board sees at contract approval. A recommendation expressing the outcome as a price is not sufficient; the paper should carry the risk allocation, the three-part test results, and the resourcing attached to retained risks. This is one additional page and it changes the nature of the decision the board is making.
Long term. Decide deliberately what client-side capability the organisation will hold. An enterprise that intends to retain risk must staff for it, and an enterprise that intends to transfer risk must accept that its counterparties will price the uncertainty and defend the price. Both are legitimate positions. What fails is transferring the risk, declining to pay the premium that a genuine transfer costs, and simultaneously reducing the client-side capability that would have detected the problem early.
Signals to Monitor
- Contingency held by suppliers rather than by you. Where every line carries a supplier's risk premium, you are paying for a transfer whose value you should test.
- The volume of contemporaneous record-keeping. One interviewee notes that with large contracts "the paper trail is vast" and claims become a contest over who kept the better record. A rising documentation effort on both sides is an early indicator that the parties are preparing for dispute rather than delivery.
- Client-side headcount falling after award. The most common and least examined consequence of a transfer decision, and the one that removes your ability to see trouble early.
- Variations concentrated in one risk category. A pattern of claims clustering on ground conditions, interfaces or specification indicates an allocation that the assigned party is unable or unwilling to hold.
- Suppliers declining to bid. A shrinking bidder pool on a repeat contract form is the market pricing your allocation. It is information, and it is usually read as a procurement problem rather than a design one.
- Disputes arriving with the relationship already broken. Where the first formal claim is a surprise to your delivery team, the commercial structure and the working relationship have been running on separate tracks.
Questions for the Leadership Team
- On our largest contract, which material risks are assigned to a party that could not absorb them if they crystallised?
- What client-side capability did we retain to manage the risks we kept — and did we fund it?
- When our last major contract went badly, did the counterparty absorb the loss, or did we end up sharing it through a claim?
- What does each party do, commercially, on the day this programme starts to fail? Have we asked?
- Are we relying anywhere on a benchmark or counterfactual that no one has verified?
- Who on our side owns the commercial structure after signature, and do they have the standing to change how it is administered?
Closing Perspective
Risk allocation is treated as a procurement outcome and is in fact one of the most consequential strategic choices an enterprise makes. It determines who is watching, who is incentivised to raise a problem early, and what everyone does when the programme goes wrong — which it will, because programmes do.
A transfer that fails the absorb test does not remove exposure. It defers it, strips out the early warning that came with proximity, and converts a manageable operational problem into a legal one, at the point of maximum cost and minimum flexibility.
The alternative is not to retain everything. It is to be honest about which of your counterparties could actually take the hit, to pay properly for the transfers that are real, and to fund the capability that retention requires. Organisations that do this look, from the outside, like they are accepting more risk than their peers. What they are actually doing is declining to pretend.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.