Design contingency exists because a drawing set at fifty percent does not contain enough information to price. It is money held against what the design has not yet decided, and it should therefore fall as the design is decided. On most projects nobody ever checks whether it did.
Three funds doing three different jobs
By the time a project reaches a guaranteed maximum price there are usually three separate pots of money, and owners routinely treat them as one.
Design contingency sits in the owner budget and covers development of the design between the current stage and completion of the documents. It exists because a schematic design becomes a construction document set and things get added along the way, not because anybody made a change.
Construction contingency sits inside the contract, usually with the contractor, and covers what goes wrong during construction with a design that is already complete.
Owner contingency sits outside the contract entirely and covers owner decisions, scope changes, and the things neither of the other two is for.
These are not interchangeable. The first is a function of design completeness, the second of construction risk, the third of the owner’s own behavior.
The commonest budget error on an owner side model is carrying one combined figure that is meant to serve all three, which produces a number that looks adequate and is exposed in all three directions at once. The separation between the second and third is set out in contractor contingency and owner contingency.
How design contingency is supposed to behave
The convention is a declining scale tied to design stage. Something in the region of fifteen percent at concept, ten at schematic, seven at design development, three to five at construction documents, and close to nothing once the documents are complete and coordinated.
The logic is that the unknown is the design itself, so as the design is resolved the unknown shrinks.
Two things follow from that logic and both are frequently ignored.
First, the reduction should be evidenced rather than assumed. Moving from ten percent to five because the project has reached design development is a calendar based reduction, not a risk based one. The question is whether the documents actually resolved what they were supposed to resolve.
Second, the money released should go somewhere deliberate. A design contingency that falls from seven percent to four has released three percent of project cost, and that is a decision about where it goes rather than an accounting entry.
On a great many projects it goes into scope. The design team, knowing the budget has headroom, develops something slightly more ambitious, and the contingency is spent on the design rather than held against it.
The handover, which is where it goes wrong
At the guaranteed maximum price the design contingency is supposed to hand over to the construction contingency. That is the moment the risk changes character, from what the drawings do not say to what the building does not do.
Three failures happen at that handover and all three are common.
Double counting. The owner retains a design contingency and the contractor carries a construction contingency covering some of the same uncertainty. Both are real money and one of them is unnecessary.
Silent transfer. The design contingency is simply absorbed into the contract value, funding scope development, and the owner arrives at signature with no design contingency and a construction contingency sized as though the documents were complete.
Premature release. The design contingency is released to the development budget at the guaranteed maximum price on the basis that the price is now fixed, when in fact the documents are at eighty percent and the remaining twenty percent will produce exactly the kind of development the contingency was for.
The third is the most damaging because it happens at the moment the owner feels most secure.
A worked example
Illustrative figures. Not taken from any client project and not a quotation.
A 63 million dollar construction budget on a healthcare project. At design development the owner model carries $4.4 million of design contingency, seven percent.
The guaranteed maximum price is agreed at eighty five percent documents. The contract includes a contractor contingency of $2.5 million, four percent.
The owner releases $3.2 million of the design contingency to the development budget at signature, retaining $1.2 million, on the reasoning that the price is now guaranteed.
Over the following eleven months, completion of the remaining fifteen percent of documents produces 62 design development items. None is a change of scope by the owner. Each is the drawings now saying something they did not say before.
Total $2.2 million. The contractor contingency covers part, the retained owner design contingency covers part, and $1.4 million is funded from elsewhere in the development budget.
The design contingency was released against a guarantee that did not cover the thing the contingency was for, because a guaranteed maximum price guarantees the price of the documented scope rather than the completeness of the documents, as in what a guaranteed maximum price actually guarantees.
The test to apply before releasing any of it
One question: what percentage of the documents is complete, and what is in the remaining percentage.
Not the stage label. The actual content. A set described as eighty five percent may have complete architectural and structural drawings with mechanical and electrical at sixty, which means the remaining work is concentrated in exactly the trades that produce coordination change.
Three checks make the answer concrete. Which specification sections are still marked as not issued or pending. Which drawing sheets are placeholders. And how many design questions are currently open and unanswered.
A set with forty open questions is not eighty five percent complete regardless of what the cover sheet says, and the design contingency should be sized against those forty rather than against a percentage.
This is the same discipline as sizing construction contingency against named events rather than as a percentage, described in how much contingency is enough, applied one stage earlier.
Who should hold what, and why it matters at signature
Where the documents are genuinely complete, the design contingency should be close to zero and the construction contingency carries the project. That is the clean case and it is uncommon.
Where the documents are at eighty to ninety percent, which is where most guaranteed maximum prices are agreed, the owner needs to retain a design contingency and needs to say so explicitly in the budget rather than treating the guarantee as covering it.
Where the documents are below eighty percent, the honest position is that the guarantee covers a smaller proportion of the eventual building than the contract value suggests, and the allowances and qualifications are doing the work the drawings should be doing.
In all three cases the useful output is one line in the budget that says what the design contingency is for, what stage it is sized against, and what evidence would justify releasing it.
That line is what prevents the release from happening by default at the moment of signature, which is when it usually happens and when it is least justified.
What we do
We read the document set for what is actually resolved rather than for the percentage on the cover, count the open questions and the pending sections, and say what design contingency the remaining work justifies. Then we check it against the contractor contingency to find the double counting and the gaps. That work is part of the readiness review. With three weeks or more before signature, the full pre-GMP review reads the price, the schedule, the interfaces and the change exposure together.
Questions people ask
Should the design contingency ever transfer to the contractor?
Rarely, because the two funds cover different risks and the contractor has no control over design development. What can transfer is a specific scope once it is defined, converted into a priced item rather than a contingency. Transferring an undefined design risk usually means paying a premium for somebody to carry uncertainty they cannot manage.
Is a percentage scale by design stage useful at all?
As a first check, yes, because it tells you whether the number is in the right region. As a basis for decision, no. Two projects at the same stage can have very different amounts of unresolved design, and the percentage cannot see the difference between them.
What if the design team says the documents are complete?
Then the open question log should be empty and no specification sections should be pending, and both are verifiable in an afternoon. Complete is a testable claim rather than an opinion, and the test costs very little compared to what a wrong answer costs eleven months later.
This is general information about construction contracts and is not legal advice.