A GMP That Lands $3.2 Million Above Budget
A 90,000 sf public library and community center has a construction budget of $48 million. At design development the estimate was $46 million in current dollars. Construction is scheduled to start in 10 months and last 20 months; the estimator applied 4 percent annual escalation from today's dollars to the construction midpoint. The CM-at-risk has now submitted a guaranteed maximum price of $51.2 million, composed of $47.5 million trade, $1.9 million general conditions, $1.3 million CM fee, and $0.5 million design contingency. The owner holds $1.5 million in unallocated contingency. Council wants a recommendation within 30 days, before the site package is bid and before the structural package is awarded.
1. Is the gap real? Reconcile the numbers before anyone starts cutting scope.
Read the worked answer
Work the arithmetic first, because part of the gap is probably double counting.
Escalation: the midpoint of construction is about 20 months out, or 1.67 years. At 4 percent compounded, 1.04^1.67 is roughly 1.068 - about 6.8 percent. So $46.0 million escalated is approximately $49.1 million. That alone exceeds the $48 million budget by $1.1 million before any scope change; escalation to a midpoint that far out was always going to eat the margin.
The GMP is $51.2 million. Against the escalated estimate, the unexplained delta is about $2.1 million. Ask four questions: (1) Does the $47.5 million trade figure already include escalation? If so, escalation has been counted twice. (2) What scope was added between design development and the GMP - reconcile by trade, in dollars, not by narrative. (3) How many bidders priced each package, and how does that compare with the estimate's market assumptions? (4) Is the $0.5 million design contingency appropriate at this stage, and where is the construction contingency?
Present the reconciliation as a table. Do not negotiate a number you have not decomposed.
2. How should contingency be structured across design, construction, and owner reserves at this point in the project?
Read the worked answer
Contingency should shrink and migrate as risk is retired, and each pool should have a named owner and a written draw protocol.
- Design contingency covers incomplete design and detailing. At GMP with documents near complete, it should be small - 1 to 2 percent, not a hidden cushion. A $0.5 million line on $47.5 million is about 1 percent, which is defensible, but confirm it is genuinely design risk and not unbid scope.
- Construction contingency covers field conditions, coordination clashes, and trade default. On a renovation-adjacent institutional project it typically runs 3 to 5 percent, and it should sit under the owner's control or be jointly released, not rolled into the CM fee.
- Owner contingency covers scope change and unknowns the owner owns. The existing $1.5 million is the right pool to bridge a gap, and spending it is a decision the council must make explicitly.
Guardrails: no contingency may be used to fund scope growth, all draws require written justification and a log, and unused contingency returns to the owner at closeout. Escalation risk after GMP signing belongs to the CM, not the owner - state that in the amendment.
3. Council wants $3.2 million removed. Where do you cut, and what do you refuse to cut?
Read the worked answer
Sequence the options from reversible to irreversible.
Use first: apply the $1.5 million owner contingency, and correct any escalation double count. That may close most of the gap with no design change at all.
Cut next (least lifecycle harm): furniture, fixtures, and equipment packages; decorative ceiling and millwork; site landscaping and paving beyond accessibility requirements; a phased fit-out of the second-floor community rooms with shell space and utilities stubbed for later.
Consider with care: reducing gross area by 5 percent, which saves roughly $2 million at $450/sf and is the only move large enough to matter, but it changes the program and the approval basis.
Refuse: envelope insulation and air-barrier continuity, glazing performance, roof drainage, and mechanical capacity. These are cheap to keep and expensive to reverse; cutting them shifts cost into thirty years of energy and moisture repairs. Also refuse to cut commissioning.
Note the cost of delay: each month adds roughly 0.33 percent, about $170,000. A 30-day decision window is itself worth money - decide, then document what was deferred and when it returns.